Gemma 3n E4B Instruct

June 26, 2025Last synced Apr 7, 2026, 4:00 PM

Blended Price

$0.025/M

Input Price

$0.020/M

Output Price

$0.040/M

Speed

36 tok/s

TTFT

0.29s

Benchmark Scores

Intelligence Index

External Benchmarks

AA-Omniscience Hallucination Rate

How Gemma 3n E4B Instruct 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

0.3 / 100

LiveCodeBench

0.1 / 100

Terminal-Bench Hard

0.0 / 100

τ²-Bench

0.1 / 100

AA-Omniscience AccuracyPredicted

43.5 / 100

Aider PolyglotPredicted

19.1 / 100

AlpacaEval 2.0 LCPredicted

30.5 / 100

AlpacaEval 2.0 RawPredicted

24.0 / 100

ARC-AGI-1Predicted

8.6 / 100

ARC-AGI-1 Cost per TaskPredicted

0.0

ARC-AGI-2Predicted

1.7 / 100

ARC-AGI-2 Cost per TaskPredicted

0.1

BFCL (Berkeley Function Calling)Predicted

30.4

BigCodeBench CompletePredicted

45.8 / 100

BigCodeBench InstructPredicted

36.8 / 100

AA Intelligence Index (Matrix)Predicted

61.3

AA Long Context Reasoning (Matrix)Predicted

72.2

AIME 2024Predicted

42.6

AIME 2025 (Matrix)Predicted

46.3

Arena-Hard AutoPredicted

51.5

BrowseCompPredicted

69.4

BRUMO 2025Predicted

91.9

CMIMC 2025Predicted

78.6

GPQA Diamond (Matrix)Predicted

51.9

HLE (Matrix)Predicted

15.9

HMMT Feb 2025Predicted

45.7

HMMT Nov 2025Predicted

90.5

HumanEvalPredicted

69.8

IFBench (Matrix)Predicted

32.2

IFEvalPredicted

78.6

IMO 2025Predicted

17.0

LiveCodeBench (Matrix)Predicted

40.7

MATH-500 (Matrix)Predicted

89.9

MathArena Apex 2025Predicted

1.8

MMLU-Pro (Matrix)Predicted

67.6

MMMU-ProPredicted

78.9

MRCR v2Predicted

76.5

OSWorldPredicted

37.5

SimpleQAPredicted

26.7

SMT 2025Predicted

83.0

SWE-bench ProPredicted

30.0

Tau-Bench Telecom (Matrix)Predicted

96.8

Terminal-Bench 2.0Predicted

31.8

Terminal-Bench 1.0Predicted

17.0

USAMO 2025Predicted

6.3

Video-MMUPredicted

85.4

browsecompPredicted

68.7

BullshitBenchPredicted

28.1 / 100

Aider PolyglotPredicted

0.0

Apex AgentsPredicted

2.7

Arc Agi 2Predicted

0.0

BIG-Bench HardPredicted

3.0

CAD-EvalPredicted

7.2

Chess PuzzlesPredicted

0.2

CyBenchPredicted

0.1

DeepResearchBenchPredicted

0.3

FictionLiveBenchPredicted

0.3

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

5.2

HellaSwagPredicted

2.9

Lech Mazur WritingPredicted

7.0

METR Time HorizonsPredicted

0.8

OTIS Mock AIME 2024–2025Predicted

0.1

PosttrainbenchPredicted

0.0

SimpleQA Verified (Epoch)Predicted

0.3

The Agent CompanyPredicted

1.1

TriviaQAPredicted

3.9

WinoGrandePredicted

0.7

FrontierMathPredicted

17.5 / 100

GAIA Level 1Predicted

19.9

GAIA Level 2Predicted

4.4

GAIA Level 3Predicted

0.0

GAIAPredicted

12.1 / 100

LegalBenchPredicted

66.5 / 100

LiveBench CodingPredicted

69.1 / 100

LiveBench Data AnalysisPredicted

47.7 / 100

LiveBench LanguagePredicted

58.7 / 100

LiveBench MathPredicted

71.8 / 100

LiveBench OverallPredicted

50.9 / 100

LiveBench ReasoningPredicted

46.8 / 100

LongBench v2 EasyPredicted

37.4

LongBench v2 HardPredicted

34.5

LongBench v2Predicted

31.4 / 100

MathVistaPredicted

53.5 / 100

MedQA (USMLE)Predicted

63.4

MLE-benchPredicted

25.6 / 100

MMMUPredicted

60.1 / 100

MMTU Table UnderstandingPredicted

48.3 / 100

MT-BenchPredicted

7.8 / 10

NoLiMa (NIAH)Predicted

89.4 / 100

OCRBench v2Predicted

72.2 / 100

Open LLM AveragePredicted

32.8 / 100

Open LLM: BBHPredicted

59.2 / 100

Open LLM: GPQAPredicted

32.3 / 100

Open LLM: IFEvalPredicted

69.9 / 100

Open LLM: MATH Level 5Predicted

27.4 / 100

Open LLM: MMLU-PROPredicted

43.3 / 100

Open LLM: MUSRPredicted

42.2 / 100

RE-BenchPredicted

12.6

SimpleBenchPredicted

25.9 / 100

simpleqaPredicted

23.2

SWE-bench LitePredicted

11.5 / 100

SWE-bench VerifiedPredicted

36.2 / 100

tau-bench RetailPredicted

81.7 / 100

Vectara Factual ConsistencyPredicted

92.4 / 100

Vectara Hallucination RatePredicted

7.6 / 100

WebArenaPredicted

1.5 / 100

WeirdMLPredicted

43.0 / 100

WildBenchPredicted

43.2

TruthfulQAtruthfulqa

68.7 / 100