Open-weight model with strong multilingual support, at a third the price of Mistral Medium 3.5 — a step below it on complex reasoning.
Context window
256K tokens
Max output
4.1K tokens
Provider
Mistral AI
Pricing
PDF support
Zero Data Retention