Journalism begins where hype ends

,,

Solve intelligence, and then use that to solve everything else."

—Demis Hassabis (DeepMind)

Mistral AI Releases Weights of Mistral Large 4 a.k.a Le Chonk- Europe’s “Heaviest” LLM

The LLM is named after a viral obese orange cat “Le Chonk”, a reference to ML 4’s 1 trillion parameter with 52 billion per token active parameters, making it one of the “heaviest” open-weight LLMs.
Image of Le chonk the cat with computer as brand image of Mistral Large 4 model of Mistral AI.
October 30, 2026 11:54 AM IST | Written by Vaibhav Jha

French AI company Mistral released the weights of their flagship AI model Mistral Large 4 a.k.a Le Chonk, a 1-trillion parameter, natively multimodal mixture-of-experts LLM , that is trained from scratch on European infrastructure.

The AI company released the weights of ML 4 on Hugging Face platform while users can also avail it through Mistral API and OpenRouter. The pricing has been kept at $0.70 input / M, $0.07 cached input / M, and $2.10 output / M tokens.

Mistral AI had earlier in October launched a public preview of Mistral Large 4 with selected partners given early access. The LLM is named after a viral obese orange cat “Le Chonk”, a reference to ML 4’s 1 trillion parameter with 52 billion per token active parameters, making it one of the “heaviest” open-weight LLMs.

 

Touted as an AI model designed for Sovereign AI European infrastructure, Mistral Large’s training data spans more than 160 languages including every official language in the European Union.

“With Mistral Large 4, that control begins with the model itself. Organizations have the ability to inspect, customize, and deploy the model on their own infrastructure. They can use and improve Mistral Large 4 without depending on Mistral to serve it through an API. Mistral Large 4 was developed and trained from scratch in Europe, giving organizations clarity about the origins of the technology they’re building on,” read a blogpost by Mistral.

Mistral Large 4 Leads in FrontierCode, Claims Mistral AI

According to Mistral AI, the ML 4 delivers frontier coding performance using significantly lesser tokens. The company claimed that on Cognition’s private FrontierCode 2.0 benchmark, Mistral Large 4 leads with a score of 19.7% on end-to-end code quality, ahead of Kimi K3 and DeepSeek V4.1 Flash.

On the Artificial Analysis Cyber Index, which measures a model’s ability to discover vulnerabilities in a codebase, Mistral Large 4 scores 51%, trailing just behind DeepSeek V4.1 Flash.

“On the Artificial Analysis Coding Agent Index, Mistral Large 4 scores 48%, effectively tied with the leading open-weight model, GLM-5.3, while using 47% fewer tokens per task on average. Our model’s built-in guardrails distinguishes between defensive and malicious, offensive cybersecurity tasks. It retains strong performance on Cyber benchmarks, while achieving an average refusal rate of 96.6% on malicious cyber requests, the highest we measure among open models,” read the post.

Also Read: Mistral Large 4: French AI Company Unveils 1 Trillion-Parameter Multimodal Model

Author

  • Vaibhav Jha, editor and co-founder at AI FrontPage

    Vaibhav Jha is an Editor and Co-founder of AI FrontPage. In his decade long career in journalism, Vaibhav has reported for publications including The Indian Express, Hindustan Times, and The New York Times, covering the intersection of technology, policy, and society. Outside work, he’s usually trying to persuade people to watch Anurag Kashyap films.

    LinkedIn

Related Posts

LATEST NEWS