Mistral-Small-4-119B-2603
Mistral AI- 119.4B parameters
- Mixture of experts, 6.5B active
- Apache-2.0
- Released in FP8
- Estimated
mistralai/Mistral-Small-4-119B-2603 on Hugging Face
Mistral-Small-4-119B-2603 with Glyd
| As released | Glyd | Change, % | ||
|---|---|---|---|---|
| Weights in GPU memory, GBFP8 weights counted as they are | 120.92 | 119.93est. | −1 | |
| Smallest single GPU, GBWeights, an 8K-token KV cache and 1.5 GB for the runtime | 141 | 141 | ||
| H100 80 GB GPUs it takes | 2 | 2 |
Which single GPU it fits
FP8 weights counted as they are.
| Memory | GPUs | As released | Glyd |
|---|---|---|---|
| 16 GB | RTX 4080, RTX 5080 | No | No |
| 24 GB | RTX 4090, RTX 3090, A10 | No | No |
| 32 GB | RTX 5090 | No | No |
| 48 GB | RTX A6000, L40S, RTX 6000 Ada | No | No |
| 80 GB | H100, A100 80 GB | No | No |
| 96 GB | RTX PRO 6000, GH200 | No | No |
| 141 GB | H200 | Fits | Fits |
Related models
Measured so far
Twenty-two open models measured, with quality wherever bf16 fits one GPU.