The LLM benchmark aggregator.

MiniMax M3

MiniMax · 427B · open weights

Strongest in Computer use (#5 of 68, on 1 of its 2 boards), weakest in Safety (#237 of 329). Above par in 21 of 36 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5 and Claude Fable 5.1 and ahead of Kimi K2.5 and Claude Sonnet 4.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Professional54.6−7.8#35/1941/2

Medical56.4−10.1#30/1491/1

Finance56.4−6#33/1621/1

Cybersecurity50.1−13.8#33/471/1

Legal52.9−13#68/1731/1

Agents57−9.4#36/2983/6

Computer use63.2−4.9#5/681/2

Knowledge work56−14.7#60/2082/2

Tool use43.9−30.1—/81✱0/1

Web research55.4−8.5—/35✱0/2

Workflow automation51.1−14.9—/0✱✱

Human preference60.1−8.7#86/3541/1

Knowledge51.6−10.9#92/1971/3

Academic knowledge53.7−10.2#63/1231/1

Factuality52.3−8.5—/0✱✱

Coding45.6−20.6#121/1792/6

Agentic coding45.4−21.5#116/1732/5

Code generation57.6−7.7—/20✱0/1

Repository Q&A42.9−23.8—/0✱✱

Reasoning43.3−22.1#151/1953/5

Science61.6−1.9#14/1221/1

Reasoning47.2−21.8#80/1471/3

Mathematics32.5−31.3#152/1522/2

Expert reasoning61.4−5—/0✱✱

Core abilities42.9−22.4#164/1992/4

Language48.3−18.3#39/621/1

Instruction following31.4−42.4#57/621/1

General intelligence46.2−23.2#115/1781/3

Long context64.6leads—/0✱✱

Safety47.6−13.8#237/3291/3

Toxicity avoidance55.5−3.5#91/2681/1

Secure code56.4−9.8#102/2701/1

Fairness48.9−21.1#127/2921/2

Harm refusal45.6−15.9#230/2921/2

Jailbreak resistance33.9−31.1#236/2681/1

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for MiniMax M3, left for the other.

Command Awon 19 · lost 3

GPT-4.1 nanowon 18 · lost 6

GPT-4o miniwon 17 · lost 6

GLM-4.6won 17 · lost 6

NVIDIA Nemotron 3 Ultra 550B A55Bwon 19 · lost 7

Llama 4 Scout Instructwon 16 · lost 6

Gemini 2.0 Flashwon 16 · lost 6

Llama 4 Maverick Instructwon 17 · lost 7

DeepSeek V3won 17 · lost 7

Claude Haiku 4.5won 15 · lost 7

GPT OSS 120Bwon 16 · lost 8

DeepSeek R1won 15 · lost 8

GPT-4owon 14 · lost 8

Grok 3won 14 · lost 8

O3 Miniwon 13 · lost 9

Claude 3.7 Sonnetwon 12 · lost 10

GPT-4.1 miniwon 13 · lost 11

GPT-5 miniwon 13 · lost 11

Grok 4.3won 13 · lost 11

O4 Miniwon 12 · lost 11

Gemini 2.5 Flashwon 12 · lost 11

Grok 4won 12 · lost 11

Kimi K2won 12 · lost 11

Qwen3 235B A22B Instructwon 12 · lost 11

Claude Sonnet 4won 13 · lost 12

Kimi K2.5won 13 · lost 12

DeepSeek V3.2won 12 · lost 12

Kimi K2.6won 12 · lost 13

Qwen3.7 Maxwon 12 · lost 13

GPT-4.1won 11 · lost 12

Inklingwon 13 · lost 15

GLM-5.2won 12 · lost 14

Qwen3.8 27Bwon 10 · lost 12

Gemini 3.6 Flashwon 11 · lost 15

O3won 10 · lost 14

Gemini 2.5 Prowon 10 · lost 14

Mistral Large 4won 9 · lost 15

DeepSeek V4 Flashwon 10 · lost 17

GLM-5.1won 8 · lost 14

Claude Sonnet 4.5won 9 · lost 16

Gemini 3.5 Flashwon 9 · lost 16

MiMo V2.6 Prowon 8 · lost 15

GPT-6 Lunawon 9 · lost 17

Claude Sonnet 4.6won 8 · lost 16

Claude Opus 4won 7 · lost 15

DeepSeek V4 Prowon 8 · lost 19

GPT-5won 7 · lost 17

GLM-5.3 Flashwon 7 · lost 20

Qwen3.8 Maxwon 7 · lost 21

GPT-5.1won 5 · lost 17

GPT-5.6 Terrawon 6 · lost 22

Gemini 3.8 Flashwon 6 · lost 22

GPT-6 Solwon 5 · lost 19

Muse Spark 1.1won 5 · lost 20

Claude Sonnet 5won 5 · lost 21

GPT-5.4won 5 · lost 21

GPT-5.6 Lunawon 5 · lost 22

Grok 4.5won 5 · lost 22

Grok 4.6won 5 · lost 23

Claude Opus 4.6won 3 · lost 19

GPT-5.2won 3 · lost 22

Claude Opus 4.8won 3 · lost 23

Claude Opus 5.5won 3 · lost 23

Kimi K3won 2 · lost 20

Grok 4.7won 2 · lost 23

GPT-5.5won 2 · lost 24

Claude Opus 4.7won 2 · lost 24

GPT-5.6 Solwon 2 · lost 25

Claude Fable 5.1won 1 · lost 21

Claude Opus 5won 0 · lost 27

§ 3 · Sources

Where the numbers come from

11 publications, 59 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-11-10

LMArena Text 1436

Artificial Analysis ↗1 measureread 2026-11-10

Artificial Analysis Intelligence Index 19

GDPval-AA ↗1 measureread 2026-11-10

GDPval-AA 1250

AA-Briefcase ↗1 measureread 2026-11-10

AA-Briefcase 1089

LiveBench 1.0 ↗8 measuresread 2026-11-10

LiveBench 1.0 67.3LiveBench 1.0 · Reasoning 74.5LiveBench 1.0 · Coding 68.2LiveBench 1.0 · Agentic Coding 40.7LiveBench 1.0 · Mathematics 76.9LiveBench 1.0 · Data Analysis 76.2LiveBench 1.0 · Language 76.8LiveBench 1.0 · Instruction Following 57.5

OSWorld ↗1 measureread 2026-11-10

OSWorld 75.2%

SimpleBench ↗1 measureread 2026-11-10

SimpleBench 45.8%

Vals.ai ↗15 measuresread 2026-11-10

Vals · Legal Research Bench 29.81%Vals · LegalBench 85.42%Vals · Finance Agent 48.27%Vals · CorpFin 68.1%Vals · TaxEval 72.73%Vals · MortgageTax 68.36%Vals · MedCode 46.29%Vals · MedScribe 87.25%Vals · CyberBench 65%Vals · SWE-bench Verified 75% (Mini-SWE-agent)Vals · Vibe Code Bench 47.57% (OpenHands)Vals · Code Migration 19.93%Vals · GPQA Diamond 92.68%Vals · MMLU Pro 84.22%Vals · ProofBench 18%

Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-11-10

Enkrypt · Jailbreak risk 26.3%Enkrypt · Harmful content risk 5.6%Enkrypt · CBRN risk 49.2%Enkrypt · Toxicity risk 2.4%Enkrypt · Bias risk 79.6%Enkrypt · Insecure code risk 20%

✱✱ IFM — K2 Horizon launch ↗the model’s own publisher18 measuresread 2026-11-03

GDPVal-AA 1380tau3-Banking 15.3%Toolathlon Verified 53.7%Automation Bench Public 20.5%Apex-Agents (pass@1) 23.8%MCPMark 48.8%BrowseComp 83.5%WildClawBench 56.4%Terminal-Bench 2.1 65.2%SciCode 45.4%SWE-Atlas-QnA 42.3%SWE Bench Pro 43.8%Humanity's Last Exam (without tools) 39%GPQA Diamond 92.9%CritPt 3.7%AA-LCR 80.3%AA-Omniscience Accuracy 17%AA-Omniscience Non-Hallucination 82%

✱ CellCog — K2 Horizon analysis ↗6 measuresread 2026-11-10

GDPVal-AA (Elo) 1380Toolathlon Verified 53.7%Terminal-Bench 2.1 65.2%SWE-bench Pro (strict) 43.8%MCPMark 48.8%SWE-Atlas-QnA (strict) 42.3%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=minimax-m3)](https://publicai.io/model-index/m/minimax-m3)