Le Chonk: Built for Coders, Trusted by Defenders.
Mistral has made the weights of this open model available, aimed at agentic programming and cybersecurity work.
What the article says
Mistral has released the weights of Mistral Large 4 (Le Chonk), its flagship 1-trillion-parameter, natively multimodal model. The open release follows a public preview launched earlier in the same month and incorporates feedback plus additional reinforcement learning. The company says it was trained from scratch on European infrastructure.
It claims gains in coding (19.7% on FrontierCode 2.0, 41% on Terminal-Bench 4.0 versus 28% in the preview) and cybersecurity (51% on the AA Cyber Index, 97.1% of attacks resisted on B3). Three reasoning modes exist: low, medium and high by default. Pricing is $0.70 per million input tokens and $2.10 per million output tokens. The model is available through the Mistral API, OpenRouter, Microsoft Foundry, or self-hosted from Hugging Face; offering it as a commercial model service requires a separate license.
Read the official article AnnuIA profile
Summary written by AnnuIA in its own words from the official page of Mistral AI. The content and rights belong to the publisher: please read the original article.