MiniMax M2: benchmarks, price and rank
MiniMax M2 from MiniMax is 115th of 123 rated AI models on Spezia's leaderboard. It costs $0.30 in, $1.20 out per million tokens.
- 115th of 123 rated models
- Rating 473
- Rated on 8 benchmarks
- Open weights
- Released Oct 27, 2025
- 1 million tokens of context
Where it is strongest
Not in the top tenth on any benchmark in use.
Where it falls behind
- 67th of 68 on SimpleBench
- 103rd of 107 on LMArena text
- 57th of 60 on Vending-Bench 2
Price: $0.30 in, $1.20 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- BrowseComp44%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
49th of 56. The maker's own number, from MiniMax, via LLM Stats.
- Humanity's Last Exam12.5%
About 2,500 questions written by experts in more than a hundred fields to be as hard as they could make them, many needing a diagram or image.
63rd of 74. The maker's own number, from MiniMax, via LLM Stats.
- Vending-Bench 2$161
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
57th of 60. Independent result from Epoch AI.
- SimpleBench25%
Multiple-choice puzzles about everyday situations, space, time and people, written to trip up models that pattern-match instead of think. Ordinary adults do well on them.
67th of 68. Independent result from Epoch AI.
- LMArena text1,340
People put a question to two anonymous models and vote for the better answer.
103rd of 107. Independent result from LMArena.
- LMArena WebDev1,299
People ask two anonymous models to build the same web page or app and vote for the better one.
89th of 98. Independent result from LMArena.
- SciCode36%
Research coding problems written by scientists across 16 fields, from physics and chemistry to biology and materials science.
21st of 24. The maker's own number, from MiniMax, via LLM Stats.
- SWE-bench Verified69.4%
Real bug reports and feature requests from popular open-source Python projects. The model has to change the code so the project's hidden tests pass.
36th of 56. The maker's own number, from MiniMax, via LLM Stats.
- GPQA Diamond78%
Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.
The maker's own number, from MiniMax, via LLM Stats.
- MMLU-Pro82%
Twelve thousand multiple-choice questions across fourteen subjects, from law and medicine to engineering, with ten options each.
33rd of 56. The maker's own number, from MiniMax, via LLM Stats.
2 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.