Check my GPU

LLMs ·

MiniMax-M1 40K

Parameters
456B (45.9B active)
Architecture
Mixture of experts
Max context
976K tokens
License
apache-2.0
Released
2025-06

How to run it

Original weights · Hugging Face

Full-precision files from MiniMax, for fine-tuning, vLLM, Transformers.

huggingface.co/MiniMaxAI/MiniMax-M1-40k (opens in a new tab)

No sign-in needed.

Links checked 6 days ago against Hugging Face, Ollama and GitHub.

Memory needed

Weights plus KV cache and runtime overhead. Leave a little extra free for the operating system and other apps.

KV cache: hybrid linear attention: only the softmax-attention layers (1 in 8) keep a KV cache.

Quantization Quality Weights Total at 8K context Total at 32K context Download
Q2_K Low 182 GB 183 GB 184 GB All files (opens in a new tab)
Q3_K_M Low 223 GB 224 GB 225 GB All files (opens in a new tab)
Q4_K_M Good 278 GB 279 GB 280 GB All files (opens in a new tab)
Q5_K_M High 324 GB 325 GB 326 GB All files (opens in a new tab)
Q6_K High 374 GB 375 GB 376 GB All files (opens in a new tab)
Q8_0 Very high 483 GB 484 GB 485 GB All files (opens in a new tab)
F16 Full 912 GB 913 GB 914 GB Original weights (opens in a new tab)

Best hardware for MiniMax-M1 40K

Grade · points Hardware Quant Memory Speed Fit
82 ptsGreat (A) · 82 of 100 points: Model quality 35.1/45, Memory fit 25.0/25, Speed 7.3/15, Memory use 15.0/15 Apple M3 Ultra (512 GB) Q4_K_M 310 GB / 384 GB 9.4 tok/s Runs well
59 ptsGood (B) · 59 of 100 points: Model quality 25.8/45, Memory fit 10.0/25, Speed 8.6/15, Memory use 15.0/15 Apple M3 Ultra (256 GB) Q2_K 199 GB / 192 GB 13.2 tok/s Runs with offload

Check it on your own hardware

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.