AgMoDB
ModelsAgentsOrchestratorsEvalsCompositesVisualizeIndustry
AgMoDB by @mistakeknot
All Models

GPT-6 Astra

Closed-weight OpenAI reasoning model for agentic, coding, and research workloads, with text and image input and text output.

September 3, 2026Last synced Nov 22, 2026, 8:17 AM1050K context128K output
Effort
  • No reasoning
  • Default
  • Low
  • Medium
  • High
  • Xhigh

Blended Price

$20.00/M

Input Price

$10.00/M

Output Price

$50.00/M

Speed

49 tok/s

TTFT

234.63s

Benchmark Scores

Intelligence Index
56.9
Coding Index
—
Math Index
—
MMLU Pro
— / 100
GPQA Diamond
96.1 / 100
HLE
54.7 / 100
LiveCodeBench
— / 100
SciCode
56.5 / 100
MATH-500
— / 100
AIME
— / 30
AIME 2025
— / 30
IFBench
— / 100
LCR
80.7 / 100
Terminal-Bench Hard
— / 100
τ²-Bench
— / 100

Performance Radar

GPT-6 Astra
Field Average

External Benchmarks

AA-Omniscience Accuracyaa_omniscience
61.6 / 100
AA-Omniscience Hallucination Rateaa_omniscience
49.8 / 100
AA-Omniscience Indexaa_omniscience
42.4 / 100
ARC-AGI-1arc_agi
86.0 / 100
ARC-AGI-1 Cost per Taskarc_agi
0.2
ARC-AGI-2arc_agi
59.6 / 100
ARC-AGI-2 Cost per Taskarc_agi
0.4
BullshitBenchbullshitbench
69.0 / 100
Arena ELO: Codingchatbot_arena
1764.0
arena_elo_documentchatbot_arena
1467.0
Chatbot Arena ELOchatbot_arena
1441.0
arena_elo_searchchatbot_arena
1231.0
arena_elo_visionchatbot_arena
1276.0
Ale Benchepoch_ai
2951.3
Apex Agentsepoch_ai
0.6
Arc Agi 2epoch_ai
0.6
BALROGepoch_ai
0.7
Blueprint Bench 2epoch_ai
0.5
Chess Puzzlesepoch_ai
0.7
Critptepoch_ai
0.2
DeepSWE (Epoch Import, Version Unverified)epoch_ai
0.7 / 100
Dtbenchepoch_ai
1.0
Ebr Benchepoch_ai
0.8
Eci Scoresepoch_ai
172.0
Epoch Capabilities Indexepoch_ai
169.2
Frontiercodeepoch_ai
0.5
Frontiermath Erdosepoch_ai
0.0
Frontiermath Tier 4 V2epoch_ai
1.0
Frontiermath Tiers 1 3 V2epoch_ai
0.9
Frontiersweepoch_ai
0.7
Furniture Assemblyepoch_ai
0.8
Gbaevalepoch_ai
0.9
Gdp Pdfepoch_ai
0.5
GSOepoch_ai
0.8
Hleepoch_ai
0.5
Lmcaepoch_ai
64.1
Mirrorcodeepoch_ai
0.6
Mystery Game Puzzlesepoch_ai
0.8
OTIS Mock AIME 2024–2025epoch_ai
1.0
Posttrainbenchepoch_ai
0.4
Proofbenchepoch_ai
1.0
Rliepoch_ai
0.2
Scicodeepoch_ai
0.5
SimpleQA Verified (Epoch)epoch_ai
0.8
Spireclimbepoch_ai
0.3
Training Computeepoch_ai
9.4e+26
Vending Bench 2epoch_ai
15514.7
WebDev Arenaepoch_ai
1800.3
LiveBench Codinglivebench
87.5 / 100
LiveBench Data Analysislivebench
83.0 / 100
LiveBench Instruction Followinglivebench
75.6 / 100
LiveBench Languagelivebench
89.4 / 100
LiveBench Mathlivebench
96.8 / 100
LiveBench Reasoninglivebench
92.7 / 100
OpenRouter Token Volumeopenrouter_rankings
1380000000000.0
OpenRouter Weekly Trendopenrouter_rankings
50.0
SimpleBenchsimplebench
83.6 / 100
Vectara Factual Consistencyvectara_hallucination
91.3 / 100
Vectara Hallucination Ratevectara_hallucination
8.7 / 100
WeirdMLweirdml
93.3 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.19112,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.66394,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"ec09fbed9f211f59ec54cca2227c3309\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5814b0938ba4a54ba31da46fe17d4380\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.92231,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.66394,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.19112,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.92231,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"d8dfdc9958a2e1502b54603b03fbc557\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"aaa10807f3dcaae3ce6c15be7539e733\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-20

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-18

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"f4d8307b32b2fefe010860bedd0c5dd8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef15ebd7644a875688eda094d716da9a\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-17

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"69aae253e532553a920bf4deb50ca495\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"5206202f381a21893de82dcc2db93ae9\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-16

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"513aaaa353972ca0baa9c95cb95e04dd\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"169653a17d103c8dbca682167ffccaa6\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-10

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"3d7dcca4fb88130b9e7ec186fd9ae23f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"1e46e6ab8c83fb1a7b19f10be398adf2\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-09

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8ea4d2292a0a8c759aa7e80a2041122d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"94fd3d68aacca744ff3c39f63a187685\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-06

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-04

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"c4baed683b2f33c40537f2338682475f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"48d1fd5bf96ae9583213f6699f12f2cc\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-11-02

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8fc20a5217998d26c40ad2306d648194\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"095b2fa20aac96121092dab07e18cc8e\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-28

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"400b32b66e8c1b37fc50ff586b033048\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9314d4bea0d9d001c69a11614dc8d93c\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-21

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2ab08f641c16bcf3c7a59c38b35fa2e3\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"b4ed41aee6b00420557f098bd05f4165\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-19

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8cf5cb76df0b13c1684b346a2012df83\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"c6e9a78d12dc070d7dab18a1f667a410\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-15

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"57cc4a7689e563a0ffb7d275eb3a889b\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"19d71f8c9a2056c0be4bb4fd4a75a325\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-14

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"e7f673a6db195dac7f2956e3b1651145\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"4d3b46c6ffe9404ca1c0ebcfb5938883\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-13

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b57a26ca4df83acaa40eadb6632c7c61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ee35c9dbbb033dc7de85467eedb85d54\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-06

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-02

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"b52a6e82589fb2c719168d69aa0bb015\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f187ceaff4d7f6156e77c1f1254aafc8\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-10-01

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"8bfe2e142f29acec39df406a6631f81a\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"92bfaa47028eae49163960f4c5a298a0\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-30

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"16793624264c3bdded076b4e26af5a90\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"e7fb37556a7353e7da435c4c7b78ca50\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-29

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"284ac3aff89a7d8a3aca179ee1e5dd61\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6e0849a05a2c572b1050e5900a372e2b\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-28

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"397d7634d348db2ae8df5180db7b475d\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"882404aedf04dbdde2b6c0d731212a4f\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-24

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"58c953e3f78e2282d0259b52304c8ef8\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"ef004f05f102510151398f73e966a509\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"2f3beadbd513a221b41d3816ff651f54\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"f1d0d3b7e367e4a7909882624b1caca0\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-22

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"c985e4804ffc840374d9852c6b60005c\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0a820529c69911b0c1a63d8b2d3dbf0f\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-21

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Medium) · effort medium · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-medium-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'medium' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-medium-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 19284.776630000004,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-high-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'high' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-high-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18816.63138,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

99.9 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

62.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-none-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 23456.663940000002,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

96.7 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (XHigh) · effort xhigh · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-xhigh-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'xhigh' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-xhigh-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 18147.191119999996,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

98.4 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Low) · effort low · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-low-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'low' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-low-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 21297.922310000005,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-04

98.0 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra - Provider Adapter (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max-provider-adapter

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra-provider-adapter",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort using the provider adapter harness.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-max-provider-adapter",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 17331.95327,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-03

98.6 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (None) · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-none

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'none' reasoning effort.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-none",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 49791.11317000001,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-03

35.2 / 100
ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · GPT-6 Astra (Max) · effort max · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:openai-gpt-6-astra-max

View run source

Source metadata
{
  "provider": "OpenAI",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "openai-gpt-6-astra",
  "modelNotes": "Evaluation conducted with 'max' reasoning effort.",
  "modelsEtag": "W/\"a8dc9c600e07efa11ca8fb622181783f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"946807139d47f4bc368591b95885dfb4\"",
  "upstreamModelId": "openai-gpt-6-astra-max",
  "environmentCount": 25,
  "modelReleaseDate": "2026-09-02T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 26097.501720000007,
  "providerModelIdPublished": false
}

Evidence observed 2026-09-03

62.7 / 100

How GPT-6 Astra Compares

Axes

Filters

$0.00 – $67.5
≥ 1.4

Providers

ai2ai21-labsalibabaanthropicarceeawsazurebaidu
openai
zai
meta
nvidia
xai
upstage
deepseek
alibaba
mistral
nous-research
aws
google
anthropic
xiaomi
stepfun
kimi
reflection
minimax
azure
inclusionai
reka-ai
cohere
bytedance_seed
inception
tencent
thinking-machines
Bubble size = Context Window
8191.00
2000000.00
Compare with other models