Loading…
Loading…
Models / Mistral-Small
1 model (24B) · ↓ 195.4K downloads a month · from $0.18/h on 2x RTX 3060 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 |
|---|---|---|---|---|
| AWQ (stelterlab)stelterlab/Mistral-Small-24B-Instruct-2501-AWQ | vLLM | 14 GB | 19.5 GB | $0.18/h 2x RTX 3060 |
| FP8 dynamic (RedHatAI)RedHatAI/Mistral-Small-24B-Instruct-2501-FP8-dynamic | vLLM | 25 GB | 31.4 GB | $0.44/h RTX 5000 Ada |
| quantized.w8a8 (RedHatAI)RedHatAI/Mistral-Small-24B-Instruct-2501-quantized.w8a8 | vLLM | 25 GB | 31.4 GB | $0.44/h RTX 5000 Ada |
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").