MiniMax M3

Across the task benchmarks we track, MiniMax M3’s strongest relative results are in coding, web research, and reasoning. On SWE-bench Verified, it scores 80.5%, ranking #2 among 7 models with reported results in our dataset.

Updated: Nov 5, 2026Data as of: Nov 2, 2026

Compare this model

🥇

Strongest results

Strongest results for MiniMax M3 in tracked evaluation cohorts
Benchmark	Measures	Score
IFBench ↗58 constraint types	Precise instruction following	82.86%#1 of 17
SWE-bench ↗Verified / published open-model reports	Coding	80.5%#2 of 7
BrowseComp ↗Published open-model reports	Web research	83.52%#3 of 8
GPQA ↗Diamond / Artificial Analysis	Reasoning	92.93%#10 of 30

Model details

Open source / open weights13 benchmarks

Specifications for MiniMax M3
Model	MiniMax M3
Agent Index v1.2	Insufficient coverage (4/7) · Preview
Creator	MiniMax (opens in a new tab)
Availability	Open source / open weights
Context size	1,000,000 tokens (opens in a new tab)
Max output	Not reported
Knowledge cutoff	Not reported
License	MINIMAX COMMUNITY LICENSE · Model weights ↗ (opens in a new tab)

Pricing and performance

Per provider and reported configuration. Prices are USD per 1M tokens.

Provider offerings for MiniMax M3
Provider / configuration	Contexttokens	InputUSD / 1M tokens	OutputUSD / 1M tokens	LatencyTTFT · seconds	Speedtokens / sec
MiniMax (opens in a new tab) MiniMax-M3	1,000,000	$0.30 (opens in a new tab)	$1.20 (opens in a new tab)	0.699 (opens in a new tab)	119.672 (opens in a new tab)

Benchmark performance

MiniMax M3 in color, alongside leading models. Latest stored reports, best effort per report date.

Reasoning

198-question Diamond subset; independent AA evaluation

GPQA Diamond ↗

01
GPT-6 Astra96.3%
02
Gemini 3.8 Flash95.3%
03
Gemini 3.1 Pro94.1%
03
GPT-5.6 Sol94.1%
03
Muse Spark 1.394.1%
10
MiniMax M3This model92.9%

Coding

Verified split; published reports

SWE-bench Verified ↗

Terminal Use

89 tasks; Terminus 2; e2b; three repeats

Terminal-Bench 2.1 ↗

01
Kimi K385%
02
GLM-5.3-Flash84.3%
03
GLM-5.383.9%
04
DeepSeek V4 Pro 081378.7%
05
GLM-5.277.9%
07
MiniMax M3This model65.2%

Web Research

Published reports; source settings retained

BrowseComp ↗

Work Automations

Private 657 tasks; 50 turns; objective completion

AutomationBench-AA ↗

01
Gemini 4 Carbon77.8%
02
Gemini 4 Argon77.5%
03
Kimi K3.1 Pro73.3%
04
MiniMax M3.1 Flash72.52%
05
Claude Sonnet 5.571.8%
42
MiniMax M3This model21.25%

Knowledge Work

220 tasks; Stirrup agent; pairwise Elo

GDPval-AA v2.1 ↗

01
Claude Opus 5.51,866 Elo
02
Kimi K3.1 Pro1,858 Elo
03
Claude Sonnet 5.51,838 Elo
04
Claude Fable 5.11,763 Elo
05
DeepSeek V4.1 Flash 10171,746 Elo
28
MiniMax M3This model1,250 Elo

Reported benchmarks

All published results in our dataset, with versions, evaluation settings, and sources.

All published benchmark reports for MiniMax M3
Benchmark	Score	Evaluation setup	Source
AA-Briefcase1.1 / combined EloCombined Elo · Higher is better	1,089.34 Elo	selection: best reported configuration per release · MiniMax-M3 · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
AutomationBench-AA1.0.6 / AA private 657Objectives completed without guardrail violations · Higher is better	21.251%	MiniMax-M3; Private 657 tasks; 50 turns; objective completion	Artificial Analysis (opens in a new tab)independent
BrowseCompPublished open-model reportsAccuracy · Higher is better	83.52%	harness comparison: provider-reported; settings differ across models · Primary full table; WebExplorer-based agent; discard history after 64K tokens	MiniMax ↗ (opens in a new tab)provider reported
GDP.pdf100 tasks / All-pass / five attemptsAccuracy · Higher is better	9.8%	selection: best reported configuration per release · MiniMax-M3 · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
GDPval-AA2.1Elo · Higher is better	1,249.95 Elo	MiniMax-M3; 220 tasks; Stirrup agent; pairwise Elo	Artificial Analysis (opens in a new tab)independent
GPQADiamond / Artificial AnalysisAccuracy · Higher is better	92.929%	MiniMax-M3; 198-question Diamond subset; independent AA evaluation	Artificial Analysis (opens in a new tab)independent
HLE-Diamond1,000 questions / no toolsAccuracy · Higher is better	16%	selection: best reported configuration per release · MiniMax-M3 · No tools	Artificial Analysis (opens in a new tab)independent
Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better	38.971%	MiniMax-M3; original HLE text-only 2,158 questions; no tools; not HLE Diamond · No tools	Artificial Analysis (opens in a new tab)independent
IFBench58 constraint typesAccuracy · Higher is better	82.857%	selection: best reported configuration per release · MiniMax-M3 · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
SWE-benchVerified / published open-model reportsResolved · Higher is better	80.5%	harness comparison: provider-reported; settings differ across models · MiniMax primary benchmark table; internal Claude Code harness; system prompt override; average of four runs	MiniMax ↗ (opens in a new tab)provider reported
Terminal-Bench2.1 / Artificial AnalysisPass@1 · Higher is better	65.169%	MiniMax-M3; 89 tasks; Terminus 2; e2b; three repeats	Artificial Analysis (opens in a new tab)independent
Terminal-Bench2.1 / published open-model reportsTask success · Higher is better	66%	harness comparison: provider-reported; settings differ across models · Terminus 2; 8 CPU / 16GB; 2h; 128K output	MiniMax ↗ (opens in a new tab)provider reported
Terminal-Bench4.0 / mini-swe-agentTask success · Higher is better	2.02%	harness: mini-swe-agent · repeats: 3 · selection: best reported configuration per release · MiniMax-M3	Artificial Analysis (opens in a new tab)independent