#mistral

Gouverner un agent IA, c'est maîtriser deux choses : le modèle qu'il utilise, et ce qu'il fait dans...

Mistral's Voxtral 3, launched October 20, 2026, is the first voice stack I've seen where the pricing...

The weights land Oct 31 with no named license. The Q4/Q2 memory math, which Macs and servers can load it, and the llama.cpp offload command.

A tool definition dumped by the OpenAI SDK failed mistral-common's conversion. A two-line fix, merged by Mistral AI on October 21, 2026.

I share my step-by-step process for fine-tuning open-source large language models like Mistral and Llama 3 for business needs, covering LoRA, dataset preparation, and deployment strategies.

This is a preview, not the guide. Mistral Large 4's open weights have not landed yet. What follows is...

Mistral Large 4 benchmark: Gauntlet 0.5462 after a x0.6 money penalty that was our scorer's bug, Forge 2.0 median 0.2195. DeepSeek V4.1 Flash beat it on Forge.

CERN is pursuing a long-term effort to create a physics-focused foundation model for high-energy...

Mistral Large 4 is 1.05T parameters with 52B active per token. The 52B sets your compute. The 1T sets...

Mistral AI's most concrete recent product development is Mistral Large 4, or ML4, a public-preview...

TL;DR: Mistral AI announced the public preview API for Mistral Large 4 ("le Chonk") on October 6,...

Mistral AI has joined a consortium led by Marlan Space and Loft Orbital to develop Altair-Next Gen, a...