Mistral AI models: Medium, Small, Devstral or Ministral
| # | AI model | Capabilities | API $/1M (in / out) | Context window | Price / mo | Output $/1M | Overview | Visit site |
|---|---|---|---|---|---|---|---|---|
| 1 |  Mistral Medium 3.5 Mistral AI |    | $1.5 / $7.5 | 256K tokens | API only | 7.5 | View | |
| 2 | Mistral Small 4 Mistral AI | | $0.15 / $0.6 | 256K tokens | API only | 0.6 | View | |
| 3 | Mistral Large 3 2512 Mistral AI | | $0.5 / $1.5 | 262.1K tokens | API only | 1.5 | View | |
| 4 | Devstral 2 2512 Mistral AI |   | $0.4 / $2 | 262.1K tokens | API only | 2 | View | |
| 5 | Ministral 3 14B 2512 Mistral AI | | $0.2 / $0.2 | 262.1K tokens | API only | 0.2 | View | |
| 6 | Ministral 3 8B 2512 Mistral AI | | $0.15 / $0.15 | 262.1K tokens | API only | 0.15 | View | |
| 7 | Ministral 3 3B 2512 Mistral AI | | $0.1 / $0.1 | 131.1K tokens | API only | 0.1 | View | |
| 8 | Codestral Mistral AI | | $0.3 / $0.9 | 128K tokens | API only | 0.9 | View | |
| 9 | Devstral 2 Mistral AI | | 256K tokens | API only | View | |||
| 10 | Magistral Medium 1.2 Mistral AI |  | 128K tokens | API only | View | |||
| 11 | Mistral Large 3 Mistral AI | | $0.5 / $1.5 | 256K tokens | API only | 1.5 | View |
Browse by need
Mistral occupies a distinct position: a frontier-adjacent European lab that publishes weights under Apache 2.0 for much of its range. Medium 3.5 leads at $1.50 input and $7.50 output with a 256K window; Large 3, Small 4 and the Ministral family fill the tiers below, down to $0.10 per 1M.
The line-up also covers specialised work that the big labs price separately — OCR billed per thousand pages, Voxtral for speech, Codestral for code, moderation at no charge. For teams with data-residency constraints, the combination of EU hosting and downloadable weights is the practical argument.
Models from this vendor
11 Mistral AI models in this rating. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.
How to choose
- Decide early whether you need open weights — most of the range ships under Apache 2.0.
- Use the specialised models rather than forcing the flagship: OCR per page, Voxtral for speech, Codestral for code.
- Check EU hosting terms if data residency is contractual rather than preferential.
- Compare Medium 3.5 against Sonnet 5 and Terra directly; the gap is narrower than the price suggests.
Common mistakes
Who this suits
Related filters
Frequently asked questions
Which Mistral models are open?
Is there a free tier?
How does Mistral price OCR?
Which Mistral model competes with Sonnet?
Is Le Chat still a thing?
Frequently searched
Rate this page
Help us improve