Qwen3 8B (Non-reasoning)

April 28, 2025Last synced Apr 7, 2026, 4:00 PM

Blended Price

$0.31/M

Input Price

$0.18/M

Output Price

$0.70/M

Speed

86 tok/s

TTFT

0.92s

Benchmark Scores

Intelligence Index

External Benchmarks

AA-Omniscience Hallucination Rate

How Qwen3 8B (Non-reasoning) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $30.0

AgMoBench Overall≥ 3.5

Providers

ai21-labsalibabaanthropicawsazurebaiducoheredeepseek

Show quadrants

anthropic

openai

google

meta

mistral

nvidia

kimi

xai

azure

deepseek

aws

cohere

baidu

ai21-labs

zai

reka-ai

xiaomi

alibaba

minimax

ibm

Bubble size = Context Window

16384.00

2000000.00

Compare with other models

GPQA Diamond

0.5 / 100

LiveCodeBench

0.2 / 100

Terminal-Bench Hard

0.0 / 100

τ²-Bench

0.2 / 100

AIME 2024benchmark_matrix

76.0

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

40.1 / 100

Aider PolyglotPredicted

20.3 / 100

AlpacaEval 2.0 LCPredicted

48.9 / 100

AlpacaEval 2.0 RawPredicted

38.2 / 100

ARC-AGI-1Predicted

2.7 / 100

ARC-AGI-1 Cost per TaskPredicted

0.0

ARC-AGI-2Predicted

0.2 / 100

ARC-AGI-2 Cost per TaskPredicted

0.0

BigCodeBench CompletePredicted

56.3 / 100

BigCodeBench InstructPredicted

46.3 / 100

AA Intelligence Index (Matrix)Predicted

66.6

AA Long Context Reasoning (Matrix)Predicted

69.8

BrowseCompPredicted

57.2

BRUMO 2025Predicted

78.4

CMIMC 2025Predicted

65.5

HLE (Matrix)Predicted

18.9

HMMT Feb 2025Predicted

34.4

HMMT Nov 2025Predicted

88.1

IMO 2025Predicted

9.9

MathArena Apex 2025Predicted

0.6

MMMU-ProPredicted

80.3

MRCR v2Predicted

73.8

OSWorldPredicted

29.7

SimpleQAPredicted

18.9

SMT 2025Predicted

75.1

SWE-bench ProPredicted

28.9

Tau-Bench Telecom (Matrix)Predicted

94.7

Terminal-Bench 2.0Predicted

16.1

Terminal-Bench 1.0Predicted

12.6

USAMO 2025Predicted

5.0

Video-MMUPredicted

80.0

browsecompPredicted

54.0

BullshitBenchPredicted

17.4 / 100

Aider PolyglotPredicted

0.0

Apex AgentsPredicted

2.1

Arc Agi 2Predicted

0.0

BIG-Bench HardPredicted

3.0

CAD-EvalPredicted

6.7

Chess PuzzlesPredicted

0.1

CyBenchPredicted

0.1

DeepResearchBenchPredicted

0.3

FictionLiveBenchPredicted

0.3

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

11.4

HellaSwagPredicted

24.6

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

1.3

TriviaQAPredicted

2.5

WinoGrandePredicted

0.8

FrontierMathPredicted

10.5 / 100

GAIA Level 1Predicted

31.7

GAIA Level 2Predicted

40.2

GAIA Level 3Predicted

0.2

GAIAPredicted

20.2 / 100

LegalBenchPredicted

83.7 / 100

LiveBench CodingPredicted

71.2 / 100

LiveBench Data AnalysisPredicted

37.5 / 100

LiveBench LanguagePredicted

46.3 / 100

LiveBench MathPredicted

75.0 / 100

LiveBench OverallPredicted

39.6 / 100

LiveBench ReasoningPredicted

31.7 / 100

LongBench v2 EasyPredicted

42.3

LongBench v2 HardPredicted

35.3

LongBench v2Predicted

35.1 / 100

MathVistaPredicted

55.0 / 100

MedQA (USMLE)Predicted

78.4

MLE-benchPredicted

14.6 / 100

MMMUPredicted

64.6 / 100

MMTU Table UnderstandingPredicted

48.2 / 100

MT-BenchPredicted

7.9 / 10

NoLiMa (NIAH)Predicted

93.9 / 100

OCRBench v2Predicted

78.7 / 100

Open LLM AveragePredicted

41.7 / 100

Open LLM: BBHPredicted

67.0 / 100

Open LLM: GPQAPredicted

35.0 / 100

Open LLM: IFEvalPredicted

74.3 / 100

Open LLM: MATH Level 5Predicted

43.2 / 100

Open LLM: MMLU-PROPredicted

52.6 / 100

Open LLM: MUSRPredicted

44.7 / 100

RE-BenchPredicted

0.2

SimpleBenchPredicted

25.0 / 100

simpleqaPredicted

18.4

SWE-bench LitePredicted

13.4 / 100

SWE-bench VerifiedPredicted

32.4 / 100

tau-bench RetailPredicted

74.1 / 100

WebArenaPredicted

16.9 / 100

WeirdMLPredicted

20.3 / 100

WildBenchPredicted

46.0

BFCL (Berkeley Function Calling)bfcl

42.6

Vectara Factual Consistencyvectara_hallucination

95.2 / 100

Vectara Hallucination Ratevectara_hallucination

4.8 / 100