Mistral AI
Mistral — all variants, VRAM & hardware
Mistral's open-weight lineup — from the iconic 7B through Nemo 12B, Codestral 22B for coding, Small 24B, and the frontier Small 4 119B.
Variant comparison
| Variant | Params | Active | VRAM @ Q4 | Architecture | |
|---|---|---|---|---|---|
| Mistral 7B Instruct v0.3 | 7B | 7B | ~4.3 GB | dense | Details → |
| Mistral Nemo 12B | 12B | 12B | ~7.3 GB | dense | Details → |
| Mistral Small 24B | 24B | 24B | ~14.6 GB | dense | Details → |
| Mistral Small 3.2 24B | 24B | 24B | ~14.6 GB | vision | Details → |
| Mistral Small 3.1 24B | 24B | 24B | ~14.6 GB | dense | Details → |
| Mistral Small 4 119B | 119B | 6.5B | ~72.6 GB | moe | Details → |
| Mistral Large 3 | 675B | 41B | ~411.8 GB | moe | Details → |
Deep-dive guides
Mistral VRAM & hardware guides
FAQ
Frequently asked questions
What is Mistral?
Mistral is Mistral AI's mistral's open-weight lineup — from the iconic 7b through nemo 12b, codestral 22b for coding, small 24b, and the frontier small 4 119b.
How many variants are in the Mistral family?
We track 7 canonical Mistral variants in our catalog, spanning 7B to 675B parameters.
Which Mistral variant should I run?
Pick the largest variant that fits on your hardware at Q4_K_M quantization. For 8 GB GPUs, the smallest dense variants. For 24 GB GPUs, the 27B dense or 35B MoE (if available). For workstation / Mac Studio, the 122B or larger MoE. Use our fit calculator on each model page for the exact verdict.
See also