AI: Mistral Debuts New Reasoning AI Models (Magistral Series)

French startup Mistral AI has officially released the Magistral series—its first AI reasoning models developed to tackle complex, multi-step tasks using transparent “chain-of-thought” logic 

Key Highlights:

Two versions available:

Magistral Small: 24B parameters, open-source, multilingual support (English, French, Spanish, Arabic, Chinese).

Magistral Medium: Enterprise-grade, optimized for logic-heavy domains like finance, law, and science.

Transparent reasoning: Users can trace each logical step the AI takes—enhancing auditability and trust.

European open-source push: Positioned as an alternative to proprietary US-centric models, reinforcing EU’s AI sovereignty goals 

Backed by strong funding: Supported by VC investors like a16z, contributing to its $1 billion+ valuation 

This launch marks a milestone for European AI—shifting focus from just bigger models to smarter, more efficient, and ethical systems.