supermarket

Vision-language

Gemini 3.6 Flash

By Google DeepMind, in the Gemini Flash family.

Released
21 Jul 2026
Status
Legacy
Context
1M
Max output
65.5K
Input per 1M tokens
$0.75
Output per 1M tokens
$3.75
Open weights
No
License
proprietary
Parameters
—
Knowledge cutoff
Mar 2026
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 & reasoning46.4242 of 4
Coding & software engineering40.5Not ranked2 of 5
Vision & multimodal understanding0.0Not ranked1 of 3
Factuality & hallucination66.2101 of 1
Human preference55.3144 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 solved43.0%Independentlow7 Aug 2026Epoch AI
Chess PuzzlesPuzzles solved40.0%Independenthigh2 Aug 2026Epoch AI
Chess PuzzlesPuzzles solved35.0%Independentminimal7 Aug 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)95.5%Third partyhigh—https://conceptualreasoning.ai/dtbench via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.4493Third partyhigh—https://conceptualreasoning.ai/lmca via Epoch AI
Mystery Game PuzzlesPuzzles solved30.0%Independenthigh5 Aug 2026Epoch AI
Mystery Game PuzzlesPuzzles solved25.0%Independentminimal5 Aug 2026Epoch AI
Mystery Game PuzzlesPuzzles solved23.0%Independentlow5 Aug 2026Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy94.1%Independenthigh2 Aug 2026Epoch AI
GPQA DiamondAccuracy86.4%Independentlow6 Aug 2026Epoch AI
GPQA DiamondAccuracy85.9%Independentminimal7 Aug 2026Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierMath Tier 4Accuracy22.0%Independenthigh2 Aug 2026Epoch AI
FrontierMath Tiers 1-3Accuracy58.9%Independenthigh2 Aug 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy94.2%Independenthigh2 Aug 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy82.2%Independentlow7 Aug 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy80.0%Independentminimal7 Aug 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
DeepSWEPass@146.7%Third partyhigh—https://deepswe.datacurve.ai/ via Epoch AI
FrontierCodeMain score (Mean@5)34.4%Third partyunknown—https://cognition.com/frontiercode via Epoch AI
WeirdML (v2)Average accuracy56.1%Third partyhigh—https://htihle.github.io/weirdml.html via Epoch AI

Professional knowledge work

Professional knowledge work results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
APEX-AgentsPass@1 (all rubric criteria met)46.9%Source not statedunknown—Epoch AI via Epoch AI

Vision & multimodal understanding

Vision & multimodal understanding results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Furniture AssemblyScore23.3%Independenthigh10 Sep 2026Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)66.2%Independenthigh27 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1536Independenthigh25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes8857Independenthigh25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1453Independenthigh13 Sep 2026Arena
Document Arena (arena.ai)Votes2975Independenthigh13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1482Independenthigh25 Sep 2026Arena
Text Arena (arena.ai)Votes33739Independenthigh25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1280Independenthigh28 Sep 2026Arena
Vision Arena (arena.ai)Votes7076Independenthigh28 Sep 2026Arena

Other Gemini Flash versions

Sources