Mistral launches Le Chonk, the new French AI that narrows the gap with ChatGPT and Claude

Mistral AI is bringing out the heavy artillery. This Tuesday, October 6, the French startup unveiled Mistral Large 4, a new artificial intelligence model nicknamed “Le Chonk.” Mistral Large 4 is described as a multimodal AI, meaning that it is capable of processing different types of content, such as images, text or audio content.In three weeks, companies and researchers will be able to download and use the model freely. For now, however, the AI is available online to all internet users.

Meet Mistral Large 4, aka Le Chonk.

• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.

Open weights release end of October.

— Mistral AI (@MistralAI) October 6, 2026

An AI that “does not completely bridge the gap”

Interviewed by Le Figaro, Mistral co-founder Guillaume Lample acknowledges that “Le Chonk” is not yet capable of competing with the other AIs on the market, including the must-have ChatGPT and Claude. He admits that it is “a model that does not completely bridge the gap with frontier models, but significantly reduces it”.

The startup’s CEO, Arthur Mensch, for his part adds that the model, however, outperforms Chinese AIs on certain fronts, such as cybersecurity.Pierre Stock, vice-president in charge of science at Mistral, also touts « state-of-the-art performance » in cybersecurity, industrial processes and finance. According to him, the model ranks « among the best open-source models in the world on aggregate metrics ».

Also read: Le Chat becomes Vibe – Mistral bets everything on AI agents

Fewer chips and more freedom

Mistral also emphasizes the resource efficiency of its artificial intelligence model, which was trained with only 4,000 Nvidia chips, or « two to three times fewer » than Chinese competitors. It is also « significantly fewer than closed-AI competitors », such as ChatGPT. For reference, a closed model, such as those from OpenAI or Anthropic, remains on its creator’s servers.

By contrast, an open model can be downloaded, installed and customized by others. In other words, a publicly available model can be used by anyone… including by hackers.Alongside the announcement, Pierre Stock assures that Mistral will work with trusted partners and governments so that the model can be used “to defend, and not to carry out malicious attacks”.

Also read: ChatGPT vs. Claude – OpenAI and Anthropic roll out cheaper and more effective AI… instead of slowing down

With this new model, Mistral is mainly looking to show that Europe, and France in particular, is still well positioned in the AI race, which remains marked by the dominance of American giants.

👉🏻 Follow the latest tech news in real time: add 01net to your Google feeds, and subscribe to our WhatsApp channel.