MiniMax M3
MiniMax M3 is strong in long context; capable in multimodal tasks; and behind the leaders in reasoning, factuality, coding, agentic tasks, instruction following, and math. Too few results yet to rate safety or multilingual tasks.
Capability profile
Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.
Too few results yet to rate Safety or Multilingual.
Price
$0.30input$1.20outputper million tokens
From Artificial Analysis · 6 providers tracked · All prices
Evidence
63results on52benchmarks
- 13 independently verified
- 19 aggregator
- 18 vendor-reported
- 13 cross-referenced
From 10 sources · latest Oct 8, 2026 · How verification works
API features
Tool callingStructured outputsReasoning
As listed by OpenRouter
Research
9 papers reference MiniMax M3MiniMax M3 benchmark results
63 results on 52 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
6.7% behind the leader1 of 3 ranked benchmarks measured
- 83.00Oct 8, 2026
24.3% behind the leader2 of 6 ranked benchmarks measured
- 85.40Oct 7, 2026
- MMMU-Pro78.55Oct 8, 2026aa_mmmu_pro
Show 3 more multimodal resultsHide 3 multimodal results
- 1254.35Aug 25, 2026
- 1246Jun 17, 2026
- 78.10Oct 7, 2026
26.7% behind the leader5 of 6 ranked benchmarks measured
- GPQA Diamond92.93Oct 8, 2026gpqa
- LiveBench · Reasoning74.48Oct 8, 2026livebench_reasoning@2026-06-25
- Humanity's Last Exam38.97Oct 8, 2026aa_hle
- 45.80Jul 10, 2026
- 3.71Oct 8, 2026
Show 3 more reasoning resultsHide 3 reasoning results
- 3.70Jul 30, 2026
- 93.00Jul 17, 2026
- 37.00Jul 13, 2026
28.3% behind the leader2 of 4 ranked benchmarks measured
- AA-Omniscience · Non-hallucination81.57Oct 8, 2026omniscienceNonHallucination
- AA-Omniscience · Accuracy16.70Oct 8, 2026omniscienceAccuracy
30.0% behind the leader8 of 10 ranked benchmarks measured
- 80.50Oct 7, 2026
- LiveBench · Coding68.20Oct 8, 2026livebench_coding@2026-06-25
- Terminal-Bench 2.165.17Oct 8, 2026terminalbenchV21
- SciCode47.11Oct 8, 2026aa_scicode
- 1482.08Jun 9, 2026
- 59.00Oct 7, 2026
- Terminal-Bench Hard42.42Oct 8, 2026aa_terminalbench_hard
- LiveBench · Agentic Coding40.66Oct 8, 2026livebench_agentic_coding@2026-06-25
Show 3 more coding resultsHide 3 coding results
- 59.00Jul 13, 2026
- 66.00Oct 7, 2026
- Terminal-Bench 2.165.00Jul 13, 2026Terminal Bench 2.1 (Terminus-2)
30.2% behind the leader7 of 7 ranked benchmarks measured
- 83.52Oct 7, 2026
- 74.20Oct 7, 2026
- 70.06Oct 7, 2026
- 37.11Oct 8, 2026
- τ-Bench V3 · Banking15.26Oct 8, 2026tauBanking
- Terminal-Bench 4.02.02Oct 8, 2026
Show 2 more agentic resultsHide 2 agentic results
- AA ApexAgents27.70Oct 7, 2026APEX-Agents
- MCP Atlas74.20Jul 13, 2026MCP-Atlas (Public Set)
31.1% behind the leader2 of 3 ranked benchmarks measured
- IFBench82.86Oct 8, 2026aa_ifbench
- LiveBench · Instruction Following57.51Oct 8, 2026livebench_instruction_following@2026-06-25
42.7% behind the leader1 of 5 ranked benchmarks measured
- LiveBench · Mathematics76.95Oct 8, 2026livebench_math@2026-06-25
Show 1 more math resultHide 1 math result
- HMMT Feb 202684.40Jul 13, 2026HMMT Feb. 2026
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- livebench_language76.84Oct 8, 2026livebench_language@2026-06-25
- livebench_data_analysis76.17Oct 8, 2026livebench_data_analysis@2026-06-25
- 1444Jul 23, 2026
- AA Intelligence29.00Jun 15, 2026Artificial Analysis Intelligence Index
- AA Intelligence29.22Oct 8, 2026aa_intelligence_index
- 1.35Oct 8, 2026
Show 18 more resultsHide 18 results
- 30.84Sep 9, 2026
- Artificial Analysis Coding Index58.57Sep 9, 2026aa_coding_index
- Claw Eval (pass@3)74.50Oct 7, 2026Claw-Eval
- 20.00Jul 13, 2026
- 48.27Oct 7, 2026
- 1390.80Jul 30, 2026
- 84.40Jul 13, 2026
- 70.02Oct 7, 2026
- 42.13Oct 7, 2026
- 42.10Jul 13, 2026
- 45.10Oct 7, 2026
- 91.60Oct 7, 2026
- 52.60Oct 7, 2026
- 37.10Oct 7, 2026
- 52.80Jul 13, 2026
- 84.60Oct 7, 2026
- τ²-Bench Telecom (AA run)88.89Oct 8, 2026aa_tau2
- τ³-Bench Banking13.00Jul 30, 2026τ³-Banking
MiniMax M3: common questions
Who makes MiniMax M3?
MiniMax M3 is made by MiniMax.
When was MiniMax M3 released?
MiniMax M3 was released on Jun 1, 2026, according to Artificial Analysis.
What is MiniMax M3 good at?
MiniMax M3 is strong in long context; capable in multimodal tasks; and behind the leaders in reasoning, factuality, coding, agentic tasks, instruction following, and math. Too few results yet to rate safety or multilingual tasks.
How much does MiniMax M3 cost?
MiniMax M3 costs $0.30 per million input tokens and $1.20 per million output tokens, according to Artificial Analysis. We track its price at 6 providers. At a mix of three input tokens to one output token, it is cheaper than 59% of the 331 priced models we track.
How many benchmarks has MiniMax M3 been tested on?
We track 63 results for MiniMax M3 on 52 benchmarks from 10 sources, 13 of them independently verified. The latest was recorded on Oct 8, 2026.
Which API features does MiniMax M3 support?
OpenRouter lists tool calling, structured outputs, and reasoning for MiniMax M3.
About this record
Where MiniMax M3's numbers come from, and every name it appears under.
- Tracked since
- Jun 1, 2026
- Newest source mention
- Aug 27, 2026
Where the results come from
Verification: 63 scores · 13 independently verified · 19 aggregator-attributed · 13 vendor cross-reference · 18 vendor-reported. How these tiers are assigned
From 10 sources on 8 sites. Artificial Analysis supplies 20 of them; the 13 independently verified results come from 5 sites. Bars are coloured by trust tier.
- artificialanalysis.ai20
- api.llm-stats.com18
- huggingface.co10
- livebench.ai7
- thinkingmachines.ai3
- datasets-server.huggingface.co2
- lmarena.ai2
- simple-bench.com1