MiniMax-M2
MiniMax- Arena score
- 1340
- Among open models
- #69of 148
- Among all on Arena
- #217
- Votes
- 6,834
Measured on MiniMax-M2, the served version, not a local build. LMArena, CC BY 4.0, published Nov 11, 2026.
Score by category
CategoryScoreOpen rankArena rankVotes
- Hard prompts, English1377#64#1981.6K
- Applied math1375#59#193289
- Software & IT1373#69#2092.4K
- Coding1371#70#2091.5K
- Medicine & health1367#62#196316
- Chinese1366#66#208452
- English1365#70#2153K
- Multi-turn1361#60#1961.1K
- Hard prompts1357#65#2073.7K
- German1354#53#172131
- Math1353#70#216315
- Science1350#72#2201.1K
- Legal & government1349#67#212432
- Expert1340#71#210297
- Business & finance1336#70#2121.3K
- French1335#64#200135
- Long prompts1332#67#2171.7K
- Russian1331#69#213497
- Instructions1329#64#2112K
- Spanish1326#70#207206
- Non-English1313#75#2283.8K
- Polish1311#57#196281
- Media & sports1298#69#2221.2K
- Excluding ties1297#71#2185K
- Writing & literature1295#76#2401.5K
- Creative writing1287#74#237891
About as strong, smaller
Compare
gemma-3-12b-itGoogle133412.2B

Mistral-Small-3.2-24B-Instruct-2506Mistral AI133824.0B

gemma-4-26B-A4B-itGoogle143425.8B

gemma-3-27b-itGoogle135827.4B
Built on it
Ways to run it
- GGUF · llama.cpp, Ollama, LM Studiounsloth/MiniMax-M2-GGUF16 builds
- MLX · Apple silicon: mlx-lm, LM Studiolmstudio-community/MiniMax-M2-MLX-8bit8 builds
- AWQ · vLLM, SGLangQuantTrio/MiniMax-M2-AWQ2 builds
- GPTQ · vLLM, SGLangModelCloud/MiniMax-M2-GPTQMODEL-W4A162 builds
Most downloaded
BenchClub runs
Not runnable yet
Runs open once a MiniMax-M2 build is registered.
More from MiniMax
See more
MiniMax-M2.51M229B

MiniMax-M2.7222.2K229B

MiniMax-M3125.6K427B

MiniMax-M2.113.9K229B
Other 80–300B models
See more
DeepSeek-V4-FlashDeepSeek1.2M291B
Inkling-SmallThinking Machines1M266B
Hy3Tencent987.2K299B

NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA874.7K124B