Qwen3 8B (Reasoning)

April 28, 2025Last synced May 28, 2026, 9:09 AM

Human Frontier

—

Human-calibrated frontier signal, backed by Arena-style preference evidence and separate from raw AgMoBench benchmark composite scores.

Blended Price

$0.37/M

Input Price

$0.11/M

Output Price

$1.15/M

Speed

61 tok/s

TTFT

1.31s

Benchmark Scores

Intelligence Index

Performance Radar

Qwen3 8B (Reasoning)

Field Average

External Benchmarks

AIME 2024benchmark_matrix

How Qwen3 8B (Reasoning) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $37.5

AgMoBench Overall≥ 2.1

Providers

ai2ai21-labsalibabaanthropicarceeawsazurebaidu

Show quadrants

mistral

meta

openai

google

anthropic

azure

nvidia

alibaba

aws

ibm

nous-research

kimi

xai

zai

deepcogito

kwaikat

xiaomi

cohere

ai21-labs

inclusionai

minimax

baidu

deepseek

prime-intellect

liquidai

stepfun

bytedance_seed

upstage

reka-ai

inception

tencent

arcee

ai2

swiss-ai-initiative

perplexity

Bubble size = Context Window

8191.00

2000000.00

Compare with other models

GPQA Diamond

0.6 / 100

LiveCodeBench

0.4 / 100

Terminal-Bench Hard

0.0 / 100

τ²-Bench

0.3 / 100

AIME 2025 (Matrix)benchmark_matrix

67.3

Arena-Hard Autobenchmark_matrix

76.0

Codeforces Ratingbenchmark_matrix

1850.0

GPQA Diamond (Matrix)benchmark_matrix

63.3

GSM8Kbenchmark_matrix

89.8

HumanEvalbenchmark_matrix

82.0

IFBench (Matrix)benchmark_matrix

35.0

IFEvalbenchmark_matrix

80.0

LiveCodeBench (Matrix)benchmark_matrix

48.0

MATH-500 (Matrix)benchmark_matrix

94.0

MMLUbenchmark_matrix

76.9

MMLU-Pro (Matrix)benchmark_matrix

56.7

AA-Omniscience AccuracyPredicted

39.1 / 100

AA-Omniscience Hallucination RatePredicted

94.2 / 100

Aider PolyglotPredicted

22.8 / 100

AlpacaEval 2.0 LCPredicted

49.9 / 100

AlpacaEval 2.0 RawPredicted

39.0 / 100

ARC-AGI-1Predicted

2.1 / 100

ARC-AGI-1 Cost per TaskPredicted

0.0

ARC-AGI-2Predicted

0.0 / 100

ARC-AGI-2 Cost per TaskPredicted

0.0

BigCodeBench CompletePredicted

59.7 / 100

BigCodeBench InstructPredicted

49.6 / 100

AA Intelligence Index (Matrix)Predicted

68.4

AA Long Context Reasoning (Matrix)Predicted

69.2

BrowseCompPredicted

53.0

BRUMO 2025Predicted

74.9

CMIMC 2025Predicted

61.3

HLE (Matrix)Predicted

18.6

HMMT Feb 2025Predicted

33.4

HMMT Nov 2025Predicted

87.4

IMO 2025Predicted

8.5

MathArena Apex 2025Predicted

0.4

MMMU-ProPredicted

80.9

MRCR v2Predicted

73.0

OSWorldPredicted

35.2

SimpleQAPredicted

17.4

SMT 2025Predicted

72.7

SWE-bench ProPredicted

34.0

Tau-Bench Telecom (Matrix)Predicted

93.8

Terminal-Bench 2.0Predicted

12.9

Terminal-Bench 1.0Predicted

11.8

USAMO 2025Predicted

4.9

Video-MMUPredicted

80.2

browsecompPredicted

49.0

BullshitBenchPredicted

9.7 / 100

Aider PolyglotPredicted

0.0

Apex AgentsPredicted

1.9

Arc Agi 2Predicted

0.0

BIG-Bench HardPredicted

3.0

CAD-EvalPredicted

6.3

Chess PuzzlesPredicted

0.1

CyBenchPredicted

0.1

DeepResearchBenchPredicted

0.3

FictionLiveBenchPredicted

0.3

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

17.8

HellaSwagPredicted

35.5

Lech Mazur WritingPredicted

7.0

METR Time HorizonsPredicted

0.2

OTIS Mock AIME 2024–2025Predicted

0.1

PosttrainbenchPredicted

0.0

SimpleQA Verified (Epoch)Predicted

0.2

The Agent CompanyPredicted

1.4

TriviaQAPredicted

2.2

WinoGrandePredicted

0.9

FrontierMathPredicted

9.0 / 100

GAIA Level 1Predicted

39.1

GAIA Level 2Predicted

43.0

GAIA Level 3Predicted

0.6

GAIAPredicted

25.6 / 100

LegalBenchPredicted

87.9 / 100

LiveBench CodingPredicted

61.8 / 100

LiveBench Data AnalysisPredicted

34.9 / 100

LiveBench LanguagePredicted

42.9 / 100

LiveBench MathPredicted

50.9 / 100

LiveBench OverallPredicted

36.7 / 100

LiveBench ReasoningPredicted

28.1 / 100

LongBench v2 EasyPredicted

44.7

LongBench v2 HardPredicted

36.3

LongBench v2Predicted

36.0 / 100

MathVistaPredicted

55.6 / 100

MedQA (USMLE)Predicted

82.6

MLE-benchPredicted

12.5 / 100

MMMUPredicted

65.4 / 100

MMTU Table UnderstandingPredicted

48.9 / 100

MT-BenchPredicted

7.9 / 10

NoLiMa (NIAH)Predicted

94.3 / 100

OCRBench v2Predicted

79.6 / 100

Open LLM AveragePredicted

43.9 / 100

Open LLM: BBHPredicted

68.9 / 100

Open LLM: GPQAPredicted

35.7 / 100

Open LLM: IFEvalPredicted

74.5 / 100

Open LLM: MATH Level 5Predicted

47.3 / 100

Open LLM: MMLU-PROPredicted

55.1 / 100

Open LLM: MUSRPredicted

45.5 / 100

RE-BenchPredicted

0.1

SimpleBenchPredicted

24.7 / 100

simpleqaPredicted

17.7

SWE-bench LitePredicted

15.2 / 100

SWE-bench VerifiedPredicted

32.8 / 100

tau-bench RetailPredicted

71.6 / 100

WebArenaPredicted

18.6 / 100

WeirdMLPredicted

22.7 / 100

WildBenchPredicted

47.1

BFCL (Berkeley Function Calling)bfcl

42.6

EQ-Bench 3eqbench

1251.0

Vectara Factual Consistencyvectara_hallucination

95.2 / 100

Vectara Hallucination Ratevectara_hallucination

4.8 / 100