Model overview
Fill-in-the-middle is what sets codestral-latest apart: it is the only entry in this catalogue declaring that capability, which editor completion needs and a chat-shaped model does not provide. At 0.3 per million input tokens and 0.9 for output it sits 40 percent below the large tier (0.5 and 1.5) on both sides, and its output rate is more than eight times cheaper than the 7.5 of the medium tier, which matters because completion traffic is high in volume and short in answers. Against the small tier it costs twice as much to send (0.15) and 1.5 times as much to receive (0.6), the premium for code-specific behaviour. Cached input drops to 0.03.
Billing conditions: Standard API price. Mistral publishes a 90% discount for cached input; cache-write prices are not published separately.
- Context window
- 0 tokens
- Knowledge cutoff
- —
- Price last verified
- First observed
Descriptions, capabilities, and tariffs come from the linked source. First-party prices take priority; OpenRouter fallbacks are labelled.