Mistral has released a new open-weight AI model with a striking name and an equally striking scale. Called Le Chonk, the model contains 1 trillion parameters and is presented as a serious challenger to leading proprietary models.
The model comes from Mistral, the French AI company that developed Le Chonk. Mistral says anyone can use and customize the model, giving developers access to an AI system that reaches a scale usually associated with the biggest models in the field.
A trillion-parameter model with an open-weight release
Le Chonk is the nickname for Mistral Large 4, Mistral’s new 1 trillion-parameter model. The company has released it as an open-weight model, meaning anyone can use and customize it. That approach puts the model in the hands of people and organizations that want to work with it directly.
The release also gives Mistral a clear position in the competition around advanced AI models. Mistral presents Le Chonk as “by far the most capable open-weight model developed outside of China,” and says it is “very, very close” to some proprietary models.
Those claims place Le Chonk in two contests at once. The first is the race among open-weight models, where Mistral says its system leads models developed outside China. The second is the broader competition with proprietary models, where Mistral says the gap has become very small.
That is a large claim for a single release, but the model’s size gives the announcement a clear focal point. Le Chonk is not described only as another model from Mistral. It is described as a 1 trillion-parameter system that can compete with leading proprietary models while remaining available for use and customization.
Mistral says the race is still open
Guillaume Lample, Mistral’s cofounder and chief scientist, framed the company’s position around continued improvement. “There are so many domains in which you can improve models,” Lample said.
That comment points to a wider view of model progress. Mistral is not presenting Le Chonk as the final step in AI development. Instead, Lample’s words describe a field with many areas where models can improve, leaving room for future advances even after a model reaches this scale.
Lample also said, “Mistral is still in the race of getting the best model.” The statement captures the company’s ambition: Le Chonk is both a major release and part of an ongoing effort to compete at the highest level.
The wording matters because Mistral’s claim reaches beyond open-weight models. The company is not only arguing that Le Chonk stands out among models developed outside China. It is also saying the system comes close to proprietary models, which have traditionally set the benchmark for the strongest AI systems.
Why Le Chonk’s release matters
Le Chonk’s combination of scale and access gives the release its importance. Mistral Large 4 contains 1 trillion parameters, yet Mistral says anyone can use and customize it. That pairing makes the model central to the company’s effort to remain part of the race for the best AI model.
The release also sharpens the meaning of Mistral’s open-weight strategy. The company has attached its strongest claim to a model that people can access and modify, rather than keeping the system limited to a proprietary offering. For Mistral, that access is part of the model’s identity.
Joel Khalili’s WIRED article, published on October 6, 2026, describes how Mistral presents the model and its place in the market. The article’s summary of Mistral’s position is direct: Le Chonk is the best open-weight model outside China and can compete with leading proprietary models.
Whether that claim holds across every comparison will shape how Le Chonk is judged. For now, Mistral has put forward a model with a 1 trillion-parameter design, open-weight access, and a stated goal of challenging the strongest proprietary systems.
Guillaume Lample’s message is just as important as the model’s size. Mistral believes there are still many areas for improvement, and it says the company remains in the race to build the best model. Le Chonk is the latest expression of that ambition.
Based on
