Mistral AI

Mistral AI is a European AI company and platform for developers and organizations that want frontier open-weight models they can customize, deploy and own, alongside ready-to-use AI assistants.

Key capabilities include:

• Frontier and specialist open-weight models, from small, cost-efficient models to large ones, plus specialist models for OCR and document AI (OCR) and audio and speech (Voxtral). Open-weight models can be self-hosted.
• Vibe (formerly Le Chat), Mistral's assistant for chat, research, everyday tasks and all-day coding in the CLI, IDE or on the web, on web and mobile with 100+ connectors.
• Mistral Studio to build, govern and improve agents and applications, and Forge to customize models around your own data.
• Flexible deployment: run on-prem, in your VPC or hosted by Mistral, with in-region inference and dedicated GPU compute.

Pricing: Freemium. Vibe Free; Pro $14.99/month (students $5.99) with $15/month in API credits; Team $24.99; Enterprise (custom, private deployments, SAML SSO, audit logs). The API is pay-per-token, for example Mistral Large at $0.5 per million input tokens and $1.5 per million output tokens, with 50% off for batch processing.

Notable for pairing open-weight, self-hostable models with enterprise-grade sovereign AI deployment options trusted by large banks, manufacturers and public-sector bodies.