Mistral releases open weights for Mistral Large 4
Original title: Mistral Large 4 (Le Chonk) is now Open Weight
The 1-trillion-parameter multimodal model adds three reasoning modes and reported defensive-security gains.
Mistral released the weights of Mistral Large 4, also called Le Chonk, after a public preview earlier this month. The company says further reinforcement learning improved coding, defensive cybersecurity, and reasoning efficiency, and that the model now offers low, medium, and high reasoning modes. Mistral reports leading or tied open-weight coding results and a 51% Artificial Analysis Cyber Index score, with a 96.6% refusal rate on malicious cyber requests. It says the model was trained from scratch on European infrastructure.
Unconfirmed: Benchmark scores and refusal rates are figures reported by Mistral, not independently confirmed here.
- Mistral Large 4 is a 1-trillion-parameter natively multimodal model.
- Three reasoning modes are low, medium, and high, with high as default.
- Mistral reports 19.7% on Cognition's private FrontierCode 2.0.
- It reports 48% on the Artificial Analysis Coding Agent Index.
- Mistral says malicious-cyber refusal averages 96.6% among open models it measured.
- Vendors
- Mistral
- AI models
- Mistral Large 4Le Chonk
From the original article
9 min read at mistral.ai
Mistral released Mistral Large 4 (Le Chonk), a 1-trillion-parameter multimodal model, now available with improved coding, cybersecurity, and reasoning efficiency. It leads open-weight benchmarks in coding and defensive security while offering three reasoning modes for cost-performance balance. Fully open-weight and European-trained, it empowers sovereign AI with customizable, self-hosted deployment options.
By kuberwastaken · mistral.ai
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.