Today, we release the weights of a new and improved Le Chonk, advancing the open frontier for agentic coding and cybersecurity.
💻 Built to build: Le Chonk ranks #1 on @cognition’s FrontierCode among open models, and sets a new benchmark for Pareto efficiency on @artificialanalysis.
🛡 Built to defend: State-of-the-art in cyberdefense, put through the paces by the world's leading cyber vendors, including @CrowdStrike, @cloudflare, @rubrik, @zscaler, and others.
Built for sovereignty: Make it yours. Download the weights on @huggingface , customize it with Mistral Forge, and deploy it wherever you need it.
Read more from our team and customers 👇:
Oct 29, 2026 · 2:21 PM UTC
Congrats Mistral on the release of Le Chonk! We made some GGUFs for folks who can run it locally 🐱 xitter.twitterapi.xyz/UnslothAI/status/21158…
Mistral Le Chonk can now run locally!
The 1 trillion parameter model runs on a Mac Studio with our 1-bit GGUF which retains 91% top-1 accuracy.
Le Chonk is the strongest Western open model to date.
GGUF: huggingface.co/unsloth/Mistr…
Guide: unsloth.ai/docs/models/mistr…
For anyone interested in learning more, deep dive on LeChonk with @PierreStock, just published last week:
youtu.be/nxAoCPsGfe8?si=PDSk…
24
5,969
being able to tune this on your own vuln history, detection rules and internal codebase without any of it leaving the network is very good.
security models should get a lot better once they’re learning from the private context every company is currently too scared to send to an external API
1
9
2,005
Congrats @MistralAI!
Everyone can run it on @opencrabs
C’est tout le paradoxe des modèles géants : les poids peuvent être ouverts, mais la capacité de les faire tourner reste concentrée chez quelques acteurs.
Mistral fait un pas important, mais le vrai test sera de voir si les entreprises peuvent construire leurs propres agents sur ces modèles, à un coût compétitif et sans dépendre des mêmes géants du cloud.
Sinon, on aura ouvert les modèles, mais pas vraiment le marché.
1
10
2,855
