AgMoDB
ModelsAgentsOrchestratorsEvalsCompositesVisualizeIndustry
AgMoDB by @mistakeknot
All Benchmarks

GAIA

GAIA measures general-purpose agent performance on realistic multi-step tasks with tools, web navigation, and file operations. It is a practical benchmark for assistant workflows beyond single-turn QA.

agentsgaia

Strengths

Tasks require real tool use and sequential planning, limiting shortcut wins from pattern matching alone; three difficulty levels expose where agents fail as complexity rises.

Caveats

Performance depends heavily on orchestration stack, tool configuration, and prompt policy; task count is smaller than mass-scale QA benchmarks, so confidence intervals are wider.

How to interpret scores

Contamination risk: lowFreshness: static

Higher is better. Accuracy percentage across all levels.

Max score: 100

Relevant use cases

AI assistantsAgentic workflowsTool useMulti-step reasoningMulti-Step Web ResearchAutonomous Information Gathering

Caveat to keep in mind

Performance depends heavily on orchestration stack, tool configuration, and prompt policy.

Leaderboard

Last synced: Nov 22, 2026, 12:18 PM
#ModelProviderScorePrice/M
1
Google: Gemini 2.5 Pro Preview 06-05
source benchpress
Evidence observed Apr 13, 2026
Google89.1$3.44
2
Sao10K: Llama 3.1 70B Hanami x1
source benchpress
Evidence observed Apr 13, 2026
Sao10k89.1$3.00
3
Grok 4.20 0309 (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
SpaceXAI88.1$3.00
4
OpenAI: GPT-5.4
source gaia
Evidence observed Apr 2, 2026
OpenAI87.3$5.63
5
OpenAI: GPT-4o
source gaia
Evidence observed Nov 10, 2026
OpenAI87.3$4.38
6
Google: Gemma 2 9B
source benchpress
Evidence observed Apr 13, 2026
Google83.3$0.04
7
Google: Gemma 2 27B
source benchpress
Evidence observed Apr 13, 2026
Google82.0$0.65
8
GPT-5.2 Codex (Xhigh)
source benchpress
Evidence observed Apr 13, 2026
OpenAI80.6$4.81
9
Gemini 2.5 Pro
source gaia
Evidence observed Nov 10, 2026
Google79.4$3.44
10
MiMo-V2-Pro
source benchpress
Evidence observed Apr 13, 2026
Xiaomi79.3$0.00
11
GLM-5-Turbo
source benchpress
Evidence observed Apr 13, 2026
Z AI77.8$0.00
12
Gemma 4 31B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Google77.7$0.00
13
GPT-5 Codex (High)
source benchpress
Evidence observed Apr 13, 2026
OpenAI77.6$3.44
14
Gemini 2.5 Pro Preview (Mar' 25)
source benchpress
Evidence observed Apr 13, 2026
Google77.5$0.00
15
Qwen3.5 122B A10B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Alibaba77.3$1.10
16
MiniMax-M2.7
source benchpress
Evidence observed Apr 13, 2026
MiniMax77.2$0.53
17
GPT-5.1 Codex (High)
source benchpress
Evidence observed Apr 13, 2026
OpenAI75.8$3.44
18
Nemotron 3 Super 120B A12B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
NVIDIA75.7$0.45
19
Claude 3.7 Sonnet (Non-reasoning)
source gaia
Evidence observed Jul 25, 2026
Anthropic74.8$6.00
20
MiMo-V2-Flash (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Xiaomi74.6$0.15
21
MiniMax-M2.5
source benchpress
Evidence observed Apr 13, 2026
MiniMax74.5$0.52
22
MiMo-V2-Flash (Feb 2026)
source benchpress
Evidence observed Apr 13, 2026
Xiaomi74.3$0.00
23
GLM-4.7 (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Z AI74.3$0.92
24
Qwen3 Max Thinking
source benchpress
Evidence observed Apr 13, 2026
Alibaba74.2$0.00
25
GPT-5.4 mini (Xhigh)
source benchpress
Evidence observed Apr 13, 2026
OpenAI74.2$1.69
26
GPT-5.4 (Xhigh)
source benchpress
Evidence observed Apr 13, 2026
OpenAI73.9$5.63
27
Gemini 3 Pro Preview (Low)
source benchpress
Evidence observed Apr 13, 2026
Google73.4$4.50
28
Qwen3.5 35B A3B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Alibaba73.4$0.69
29
Google: Gemini 3 Pro Preview
source benchpress
Evidence observed Apr 13, 2026
Google72.6$4.50
30
Qwen3.5 27B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Alibaba72.0$0.82
31
Gemini 2.5 Pro Preview (May' 25)
source benchpress
Evidence observed Apr 13, 2026
Google71.8$3.44
32
KAT Coder Pro V2
source benchpress
Evidence observed Apr 13, 2026
KwaiKAT71.7$0.00
33
Grok 4.20 0309 (Non-reasoning)
source benchpress
Evidence observed Apr 13, 2026
SpaceXAI71.1$3.00
34
Gemini 3.1 Pro Preview
source benchpress
Evidence observed Apr 13, 2026
Google71.0$4.50
35
MiniMax-M2.1
source benchpress
Evidence observed Apr 13, 2026
MiniMax70.9$0.52
36
Step 3.5 Flash
source benchpress
Evidence observed Apr 13, 2026
StepFun70.8$0.15
37
KAT-Coder-Pro V1
source benchpress
Evidence observed Apr 13, 2026
KwaiKAT70.5$0.00
38
GLM 5V Turbo (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Z AI70.4$0.00
39
Google: Gemini 2.5 Pro Preview 05-06
source benchpress
Evidence observed Apr 13, 2026
Google70.4$3.44
40
Gemini 3.1 Flash-Lite
source benchpress
Evidence observed Apr 13, 2026
Google70.3$0.56
41
GPT-5.4 nano (Xhigh)
source benchpress
Evidence observed Apr 13, 2026
OpenAI70.2$0.46
42
GPT-5 mini (High)
source benchpress
Evidence observed Apr 13, 2026
OpenAI70.2$0.69
43
GPT-4.1
source gaia
Evidence observed Nov 10, 2026
OpenAI69.7$3.50
44
GPT-5.1 Codex mini (High)
source benchpress
Evidence observed Apr 13, 2026
OpenAI69.6$0.69
45
GLM-5 (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Z AI69.4$1.55
46
Gemini 3 Flash Preview (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Google69.2$1.13
47
GPT-5.3 Codex (Xhigh)
source benchpress
Evidence observed Apr 13, 2026
OpenAI68.9$4.81
48
Qwen3.5 397B A17B (Reasoning)
source benchpress
Evidence observed Apr 13, 2026
Alibaba68.7$1.35
49
Qwen Chat 14B
source benchpress
Evidence observed Apr 13, 2026
Alibaba68.4$0.00
50
Nova 2.0 Pro Preview (Medium)
source benchpress
Evidence observed Apr 13, 2026
Amazon68.0$3.44