Loading…
Loading…
Models / MiniMax-M3
1 model (~427B) · ↓ 359K downloads a month · from $5.33/h on 8x Nvidia A100 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalMiniMaxAI/MiniMax-M3 | vLLM | 854 GB | 952.4 GB | $44.03/h 8x Nvidia H200 |
| MXFP8MiniMaxAI/MiniMax-M3-MXFP8 | vLLM | 444 GB | 496.4 GB | $11.77/h 8x RTX PRO 6000 |
| MXFP4 (amd)amd/MiniMax-M3-MXFP4 | vLLM | 243 GB | 273 GB | $5.33/h 8x Nvidia A100 |
| NVFP4 (nvidia)nvidia/MiniMax-M3-NVFP4 | vLLM | 250 GB | 281.2 GB | $5.33/h 8x Nvidia A100 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").