supermarket

Vision-language

Gemma 4 31B

By Google DeepMind, in the Gemma family.

Released
31 Mar 2026
Status
Current
Context
262.1K
Max output
—
Input per 1M tokens
—
Output per 1M tokens
—
Open weights
Yes
License
apache-2.0
Parameters
30.7B
Knowledge cutoff
Jan 2025
Reasoning
Yes
Modalities
text, image, video → text

Category scores

Out of 100, from independent and third-party results only, by method category-2026-09-29.1.

Category scores
CategoryScoreRankCoverage
Knowledge & reasoning23.1732 of 4
Factuality & hallucination10.4681 of 1
Human preference47.5334 of 5

Benchmark results

Every result we hold for this model, with who ran it and where it was published.

Knowledge & reasoning

Knowledge & reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Chess PuzzlesPuzzles solved5.0%Independentminimal6 Aug 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)82.7%Third party——https://conceptualreasoning.ai/dtbench via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.3925Third party——https://conceptualreasoning.ai/lmca via Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy75.8%Independentminimal6 Aug 2026Epoch AI
Surface Evolver BenchMean score (partial credit)30.6%Third party——https://yhenon.github.io/surface-evolver-llm-eval/ via Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
OTIS Mock AIME 2024-2025Accuracy73.3%Independentminimal6 Aug 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
WeirdML (v2)Average accuracy52.3%Third party——https://htihle.github.io/weirdml.html via Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)10.4%Independent—27 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1363Independent—25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes12479Independent—25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1445Independent—13 Sep 2026Arena
Document Arena (arena.ai)Votes12326Independent—13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1453Independent—25 Sep 2026Arena
Text Arena (arena.ai)Votes6135Independent—25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1263Independent—28 Sep 2026Arena
Vision Arena (arena.ai)Votes38935Independent—28 Sep 2026Arena

Other Gemma versions

Sources