Mistral models: prices and the hardware each needs
15 models, 7 sold per token by 5 vendors. Vendors are in alphabetical order; the price is the cheapest list price among them, per million tokens.
| Model | Size | Sold per token by | From, input | Output | GPU memory |
|---|---|---|---|---|---|
| Codestral 25.08 | closed | Mistral | $0.30 | $0.90 | |
| Devstral-2-123B-Instruct-2512 | 125.03 B | not sold per token | 156.3 GB | ||
| Magistral-Small-2509 | 24.01 B | not sold per token | 60 GB | ||
| Mistral 7B | 7 B | AWS, Snowflake | $0.15 | $0.20 | 18.1 GB |
| Mistral Large | closed | AWS | $0.50 | $1.50 | |
| Mistral Large 3 | 675 B | Snowflake | $0.60 | $1.80 | |
| Mistral Large 4 | closed | Mistral | $0.70 | $2.10 | |
| Mistral Small | 24 B | AWS, Core42, DeepInfra | $0.05 | $0.08 | 58.9 GB |
| Mistral Small 4 | 119.4 B | Mistral | $0.15 | $0.60 | 149.2 GB |
| Mistral-7B-Instruct-v0.1 | 7.24 B | not sold per token | 18.1 GB | ||
| Mistral-7B-Instruct-v0.2 | 7.24 B | not sold per token | 18.1 GB | ||
| Mistral-7B-v0.1 | 7.24 B | not sold per token | 18.1 GB | ||
| Mistral-Medium-3.5-128B | 127.7 B | not sold per token | 159.6 GB | ||
| Mistral-Nemo-Instruct-2407 | 12.25 B | not sold per token | 30.6 GB | ||
| Mixtral-8x7B-Instruct-v0.1 | 46.7 B | not sold per token | 116.8 GB |