Released
1 Jun 2026
Status
Current
Context
1M
Max output
—
Input per 1M tokens
$0.30
Output per 1M tokens
$1.20
Open weights
Yes
License
minimax-community
Parameters
428B, 23B active
Knowledge cutoff
—
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 & reasoning28.0633 of 4
Coding & software engineering14.7Not ranked1 of 5
Human preference57.4103 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 solved14.0%Independent—10 Aug 2026Epoch AI
Chess PuzzlesPuzzles solved4.0%Independentnone10 Aug 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)78.9%Third party——https://conceptualreasoning.ai/dtbench via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.337Third party——https://conceptualreasoning.ai/lmca via Epoch AI
Mystery Game PuzzlesPuzzles solved8.0%Independentnone27 Aug 2026Epoch AI
Mystery Game PuzzlesPuzzles solved7.0%Independent—27 Aug 2026Epoch AI
SimpleBenchScore (AVG@5)45.8%Third party——https://lmcouncil.ai/benchmarks via Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy90.9%Independent—10 Aug 2026Epoch AI
GPQA DiamondAccuracy81.3%Independentnone10 Aug 2026Epoch AI
Surface Evolver BenchMean score (partial credit)55.0%Third party——https://yhenon.github.io/surface-evolver-llm-eval/ via Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
OTIS Mock AIME 2024-2025Accuracy71.1%Independent—10 Aug 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy26.7%Independentnone10 Aug 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierCodeMain score (Mean@5)14.7%Third party——https://cognition.com/frontiercode via Epoch AI

Computer & browser use

Computer & browser use results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
OSWorld 2.0Binary completion (500 steps)4.6%Third party——https://osworld-v2.xlang.ai/ via Epoch AI

Professional knowledge work

Professional knowledge work results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
APEX-AgentsPass@1 (all rubric criteria met)37.7%Source not stated——Epoch AI via Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1483Independent—25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes16349Independent—25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1433Independent—13 Sep 2026Arena
Document Arena (arena.ai)Votes6314Independent—13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1440Independent—25 Sep 2026Arena
Text Arena (arena.ai)Votes56646Independent—25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1237Independent—28 Sep 2026Arena
Vision Arena (arena.ai)Votes17970Independent—28 Sep 2026Arena

Other MiniMax M versions

Sources