Mistral Large, released by Paris-based Mistral AI in February 2024, is the company's flagship model and the first European LLM to compete directly with GPT-4.
**Performance**
Mistral Large scores 81.2% on MMLU, placing it between GPT-3.5 and GPT-4. On the MATH benchmark it scores 45%, and on coding benchmarks it performs comparably to GPT-4 Turbo for most practical tasks.
**Multilingual**
Natively fluent in English, French, Spanish, German, and Italian — genuine multilingual capability rather than translation-based, making it the strongest European-language model available.
**Function Calling**
Precise function calling capability makes Mistral Large well-suited for agentic applications and tool-use workflows.
**Context Window**
32K token context window — smaller than GPT-4 Turbo or Claude, but sufficient for most business applications.
**Data Sovereignty**
For European companies concerned about GDPR and data residency, Mistral offers deployment options within EU infrastructure — a significant compliance advantage.
**Pricing**
$8 per million input tokens and $24 per million output tokens via La Plateforme API. Also available on Azure and AWS.
**Verdict**
Mistral Large is the obvious choice for European businesses needing GPT-4-level performance with EU data compliance. For global users, the pricing is competitive.