Gemini 4 Argon
Claude Opus 5.5
Muse Wonder
MiMo-V2.6-Pro
Qwen3.8-Max
Kimi K3
GLM-5.3
GPT-5.5
ERNIE 5.1
DeepSeek-V4.1-Flash-20260910
Grok 4.20 Beta
amazon-nova-experimental-chat
Dola Seed 2.0 Pro
NVIDIA-Nemotron-3-Ultra-550B-A55B
Gemma 4 31B
Step 5 Preview
MiniMax-M2.5
MiniMax-M2.5 is an open-weight model from MiniMax, released Feb 12, 2026. It ranks #200 of 351 on the LMArena text leaderboard with a rating of 1359. It costs $0.30 per million input tokens and $1.20 per million output tokens, with a 200K-token context window.
- Developer
- MiniMax
- Released
- Feb 12, 2026
- Context window
- 200K
- Price in / out, per 1M tokens
- $0.30 / $1.20
- Weights
- Open · Modified MIT
- Inputs
- Text
- Reasoning
- Yes
- Tool use
- Yes
- Hugging Face downloads, 30 days
- 992.9K
- Hugging Face likes
- 1.5K
The rating LMArena published at each update since Feb 16, 2026. Ratings move as votes come in and new models join.
Benchmarks
| Benchmark | Score bar | Accuracy | Rank | Cost / test |
|---|---|---|---|---|
| SWE-benchCoding | 74.20% | 46 / 85 | $0.46 | |
| Vibe Code BenchCoding | 14.85% | 90 / 115 | $2.20 | |
| LiveCodeBenchCoding | 79.21% | 73 / 143 | $0.3 / $1.2 | |
| MedQAMedicine | 92.53% | 30 / 92 | $0.3 / $1.2 | |
| GPQA DiamondReasoning | 82.07% | 62 / 138 | $0.3 / $1.2 | |
| MMLU-ProReasoning | 80.09% | 91 / 138 | $0.3 / $1.2 | |
| AIMEReasoning | 88.75% | 29 / 96 | $0.3 / $1.2 |
Vals AI, as of Oct 24, 2026.
LMArena, by category
| Leaderboard | Rating | Rank | Votes |
|---|---|---|---|
| TextLMArena | 1359 | 200 / 351 | 42.8K |
| CodingLMArena | 1380 | 193 / 346 | 11.3K |
| MathLMArena | 1378 | 186 / 339 | 2.6K |
| Creative writingLMArena | 1331 | 193 / 349 | 6.7K |
| Instruction followingLMArena | 1353 | 186 / 351 | 13.8K |
| Hard promptsLMArena | 1372 | 187 / 351 | 26.9K |
| Multi-turnLMArena | 1364 | 191 / 349 | 7.7K |
| Long queriesLMArena | 1366 | 186 / 330 | 16.3K |
| ExpertLMArena | 1379 | 176 / 305 | 3.8K |
| Legal and governmentIndustries | 1366 | 194 / 325 | 3.3K |
| Medicine and healthcareIndustries | 1357 | 203 / 325 | 3.2K |
| Business and financeIndustries | 1357 | 188 / 344 | 8.4K |
| ChineseLanguages | 1393 | 186 / 328 | 2.9K |
| SpanishLanguages | 1353 | 193 / 249 | 1.5K |
| GermanLanguages | 1361 | 162 / 261 | 644 |
| FrenchLanguages | 1363 | 185 / 252 | 1.3K |
| JapaneseLanguages | 1172 | 225 / 236 | 368 |
| WebDevLMArena | 1386 | 105 / 124 | 7.6K |
LMArena, CC BY 4.0, as of Oct 22, 2026. Where a model is rated at several effort levels, its best is shown.
Compare MiniMax-M2.5
MiniMax
Builds long-context text, voice and video models, with open-weight releases. Founded 2021, based in Shanghai. minimax.io
Questions
Who makes MiniMax-M2.5?
MiniMax-M2.5 is made by MiniMax, founded 2021 and based in Shanghai.
How much does MiniMax-M2.5 cost?
$0.30 per million input tokens and $1.20 per million output tokens, at its provider’s list price.
What is MiniMax-M2.5’s context window?
200K tokens (204,800).
How does MiniMax-M2.5 rank?
#200 of 351 on LMArena’s text leaderboard, rated 1359 (95% interval 1355–1363) from 42,791 votes, as of Oct 22, 2026.
Is MiniMax-M2.5 open source?
MiniMax-M2.5 has open weights under the Modified MIT license, downloadable from Hugging Face (MiniMaxAI/MiniMax-M2.5).