MiniMax
MiniMax
Language model · China ·Oct 2025

MiniMax-M2

Language Open (Restricted)
229B
Parameters
18
Benchmarks

About MiniMax-M2

MiniMax-M2 is an AI model developed by MiniMax, in the language category, released in 2025, made available as an open-weights model with 229B params.

On this page you'll find MiniMax-M2's full specifications, including 18 benchmark results. Review provider pricing and benchmark scores below, or compare MiniMax-M2 head-to-head with other language models.

→ More models by MiniMax → All language models → AI benchmarks leaderboard

Links

Announcement ↗ Hugging Face ↗

Run MiniMax-M2 locally — what to buy

Share this build ↗

Sized to what MiniMax-M2 actually needs (~8k context at Q4), not to the biggest GPU: Good = cheapest that runs it, Better = best value, Best = most headroom. Speed is a hardware estimate; anything past ~120 tok/s is shown as instant.

Best
Apple M2 Ultra 192GB
Apple · ~$6,600
~5.1 tok/s
Q4 · fits full context · up to 8k ctx
Runs with: ExLlamaV2, KoboldCpp, LM Studio
See full estimate →

Software support

·ExLlamaV2 ·KoboldCpp ·LM Studio ·MLX ·Ollama ·SGLang ·TensorRT-LLM ·llama.cpp ·vLLM ?KTransformers ?text-generation-webui

✓ measured · · compatible · — not supported. Informational only — speed is hardware-based.

Apple
5.1 tok/s
1 option →
Datacenter GPU
31.8–54.3 tok/s
3 options →

Community

Would you run MiniMax-M2 again?
No votes yet — be the first.
Did it run? — community reports by hardware

No reports yet. Own this model on some hardware? Be the first to confirm it runs.

Discussion

No comments yet. Start the discussion.

MiniMax-M2 benchmark scores▶
Compare on the benchmark table →
Benchmark Category Score Variant
BrowseComp agentic 44% cited-deepseek-v3-2
BrowseComp-Zh agentic 48.5 cited-deepseek-v3-2
MCP-Mark agentic 24.4 cited-deepseek-v3-2
MCP-Universe agentic 29.4 cited-deepseek-v3-2
Terminal-Bench 2.0 agentic 30% cited-devstral-readme
Terminal-Bench 2.0 agentic 30% cited-deepseek-v3-2
Tool-Decathlon agentic 16 cited-deepseek-v3-2
τ²-Bench (TAU2) agentic 76.9 cited-deepseek-v3-2
LiveCodeBench code 83% cited-deepseek-v3-2
SWE-bench Verified code 69.4% cited-deepseek-v3-2
SWE-bench Verified code 69.4% cited-devstral-readme
SWE-Bench Multilingual coding 56.5 cited-devstral-readme
SWE-Bench Multilingual coding 56.5 cited-deepseek-v3-2
MMLU-Pro general 82% cited-deepseek-v3-2
AIME 2025 math 78.3% cited-deepseek-v3-2
GPQA Diamond reasoning 77.7% cited-deepseek-v3-2
Humanity's Last Exam reasoning 31.8% search-agent-cited-deepseek-v3-2
Humanity's Last Exam reasoning 12.5% text-only-cited-deepseek-v3-2

More models like MiniMax-M2

AI
DMind-1
 AI
COLLECTIVE-1
 AI
Cogito 1 series
 AI
Euler (Euler大模型)
· 2.5B
 AI
SmolLM-1.7B
· 1.7B
 AI
Llama-3-70B-Synthia-v3.5
· 70B

Frequently asked questions

What is MiniMax-M2? ▶

MiniMax-M2 is an AI model developed by MiniMax, in the language category, released in 2025. It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.

Who created MiniMax-M2? ▶

MiniMax-M2 was developed by MiniMax and released in 2025.

Is MiniMax-M2 open source or proprietary? ▶

MiniMax-M2 ships as an open-weights model: you can download and self-host the weights, though the licence may place some restrictions on use.

How much does MiniMax-M2 cost? ▶

Pricing for MiniMax-M2 depends on the provider. See the providers table on this page for the latest API rates.

How does MiniMax-M2 perform on benchmarks? ▶

MiniMax-M2 is benchmarked across 18 evaluations on GenAIList, including BrowseComp (44%). See the full benchmark table below and compare it with other models.

MiniMax-M2 (Tested) : Better Than GLM 4.6 (Compact & High-Efficiency AI Model)
Minimax M2.1 : RIP Gemini 3.0 FLASH! This is the BEST SMALL & CHEAP OPEN MODEL YET!
Minimax M2 (Fully Tested): I am switching to this. Better than Claude & GLM-4.6 on Long Running Task
MiniMax M2.1 Is Coming! Open-Source Agent AI Gets Better
This Model Just Changed AI Coding Forever (MiniMax M2)

Know a great video about MiniMax-M2?