Mistral Large 3 vs Mistral Medium 3.5
Facts only. This page does not pick a winner. For a picked winner, see the use case comparison on Tools.
| Mistral Large 3 | Mistral Medium 3.5 | |
|---|---|---|
| API model IDs | mistral-large-2512 | mistral-medium-2604 |
| Context window | 256,000 | 256,000 |
| Lifecycle | active | active |
| License | open-weights | open-weights |
| Trains on API data | Yes | Yes |
| Zero data retention | on-request | on-request |
| Input per 1M tokens | $0.5 (verified 2026-09-08) | $1.5 (verified 2026-09-08) |
Frequently asked questions
- Which has the larger context window, Mistral Large 3 or Mistral Medium 3.5?
- Mistral Large 3 holds 256,000 tokens and Mistral Medium 3.5 holds 256,000. Both hold 256,000 tokens. The context window counts the prompt, any attached files, and the reply together.
- Which is cheaper, Mistral Large 3 or Mistral Medium 3.5?
- Per million input tokens, Mistral Large 3 was $0.5 (verified 2026-09-08) and Mistral Medium 3.5 was $1.5 (verified 2026-09-08). Mistral Large 3 is cheaper per input token. Output tokens cost $1.5 and $7.5 respectively, and output volume usually decides the bill.
- Does Mistral Large 3 or Mistral Medium 3.5 train on API data?
- Mistral Large 3 trains on API data by default, and Mistral Medium 3.5 trains on API data by default. Zero data retention is on-request for Mistral Large 3 and on-request for Mistral Medium 3.5. Verified 2026-09-08 and 2026-09-08 against the providers' own documentation.
- Is Mistral Large 3 or Mistral Medium 3.5 being retired?
- Mistral Large 3 is active. Mistral Medium 3.5 is active. Lifecycle checked 2026-09-08 and 2026-09-08. Pin a dated snapshot ID rather than an alias if a retirement would break you.