MiniMax-M3: hardware requirements

MiniMax-M3 needs about 265 GB of memory at Q4 quantization — a 261 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.

Parameters
427B
Tier
Server only
Licence
other
Released
2026-06-02

This is a mixture-of-experts model. Only a fraction of its 427B parameters are used for each token, so it runs faster than its size suggests — but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →

Will it run on your machine?

Download sizes by quantization

Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?

QuantDownloadRAM @ 4KRAM @ 32KSource
Q4_K_Mrecommended261 GB265 GB269 GBmeasured
Q5_K_M305 GB309 GB313 GBmeasured
Q6_K369 GB373 GB377 GBmeasured
Q8_0454 GB457 GB461 GBmeasured
BF16854 GB858 GB861 GBestimated

“Measured” sizes are read from published GGUF files (bartowski/MiniMax-M3-GGUF). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.