supermarket

Vision-language

Gemini 3.1 Pro

By Google DeepMind, in the Gemini Pro family.

Released
19 Feb 2026
Status
Current
Context
1M
Max output
65.5K
Input per 1M tokens
$2.00
Output per 1M tokens
$12.00
Open weights
No
License
proprietary
Parameters
—
Knowledge cutoff
Jan 2025
Reasoning
Yes
Modalities
text, image, video, audio, pdf → text

Category scores

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

Category scores
CategoryScoreRankCoverage
Knowledge & reasoning51.4144 of 4
Coding & software engineering14.493 of 5
Agentic tool use52.412 of 2
Vision & multimodal understanding0.0Not ranked1 of 3
Factuality & hallucination73.521 of 1
Human preference52.6224 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 solved55.0%Independent—19 Feb 2026Epoch AI
Chess PuzzlesPuzzles solved49.0%Independenthigh6 Aug 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)97.1%Third partymedium—https://conceptualreasoning.ai/dtbench via Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)95.5%Third partyhigh—https://conceptualreasoning.ai/dtbench via Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)94.9%Third partylow—https://conceptualreasoning.ai/dtbench via Epoch AI
Earthborne Rangers (EBR-bench)Objectives reached14.3%Independent—25 Jun 2026Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.5382Third partyhigh—https://conceptualreasoning.ai/lmca via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.528Third partymedium—https://conceptualreasoning.ai/lmca via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.5118Third partylow—https://conceptualreasoning.ai/lmca via Epoch AI
Mystery Game PuzzlesPuzzles solved34.0%Independenthigh27 Jul 2026Epoch AI
Mystery Game PuzzlesPuzzles solved32.0%Independentmedium6 Aug 2026Epoch AI
Mystery Game PuzzlesPuzzles solved29.0%Independentlow6 Aug 2026Epoch AI
SimpleBenchScore (AVG@5)79.6%Third party——https://lmcouncil.ai/benchmarks via Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy94.4%Independenthigh6 Aug 2026Epoch AI
GPQA DiamondAccuracy94.1%Independent—20 Feb 2026Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierMath Tier 4Accuracy26.8%Independent—11 Jun 2026Epoch AI
FrontierMath Tier 4 (v1)Accuracy16.7%Independent—19 Feb 2026Epoch AI
FrontierMath Tiers 1-3Accuracy59.6%Independent—11 Jun 2026Epoch AI
FrontierMath Tiers 1-3 (v1)Accuracy36.9%Independent—19 Feb 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy95.6%Independent—20 Feb 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy95.6%Independenthigh6 Aug 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
DeepSWEPass@111.7%Third party——https://deepswe.datacurve.ai/ via Epoch AI
ExploitBenchMean capability26.1%Source not stated——Epoch AI via Epoch AI
ExploitBenchMean capability15.7%Source not stated——Epoch AI via Epoch AI
GSOOPT@122.6%Third party——https://gso-bench.github.io/index.html via Epoch AI
METR Time HorizonsAverage task success (Epoch's column)77.0%Third party——https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/ via Epoch AI
MirrorCodeScore8.9%Independenthigh10 Aug 2026Epoch AI
PostTrainBenchAverage score22.0%Source not stated——Epoch AI via Epoch AI
SWE-bench Verified% resolved75.6%Independent—24 Feb 2026Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)80.2%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)78.4%Third party——Terminal-Bench v2 Leaderboard via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)78.4%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)74.8%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)59.4%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
WeirdML (v2)Average accuracy72.1%Third party——https://htihle.github.io/weirdml.html via Epoch AI

Agentic tool use

Agentic tool use results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
BALROGAverage progress57.0%Third party——Balrog Leaderboard via Epoch AI
BALROGAverage progress56.9%Third party——Balrog Leaderboard via Epoch AI
Deep Research Bench (FutureSearch)Average score47.8%Third partyhigh—https://evals.futuresearch.ai/#drb via Epoch AI
Deep Research Bench (FutureSearch)Average score44.5%Third partylow—https://evals.futuresearch.ai/#drb via Epoch AI

Professional knowledge work

Professional knowledge work results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
APEX-AgentsPass@1 (all rubric criteria met)35.3%Source not stated——Epoch AI via Epoch AI

Long context

Long context results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
CL-benchTasks solved20.8%Source not stated——Epoch AI via Epoch AI
CL-bench LifeTasks solved16.9%Source not stated——Epoch AI via Epoch AI

Vision & multimodal understanding

Vision & multimodal understanding results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Furniture AssemblyScore26.7%Independenthigh10 Sep 2026Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)73.5%Independenthigh10 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1446Independent—25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes24363Independent—25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1459Independent—13 Sep 2026Arena
Document Arena (arena.ai)Votes49457Independent—13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1487Independent—25 Sep 2026Arena
Text Arena (arena.ai)Votes119196Independent—25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1280Independent—28 Sep 2026Arena
Vision Arena (arena.ai)Votes45204Independent—28 Sep 2026Arena

Other Gemini Pro versions

Sources