Mistral Claims 'Le Chonk' Is the Top Open-Weight Model Outside China

French AI lab Mistral has released a one-trillion-parameter open-weight model, Mistral Large 4 — nicknamed Le Chonk — and claims it is by far the most capable open-weight model developed outside of China. The model is available now in preview, with a final version due by the end of the month. Anyone can use and customize it.

The release lands at a moment when access to frontier-grade AI has become a geopolitical flashpoint. In June, the Trump administration placed temporary restrictions on distributing models from OpenAI and Anthropic, citing concerns they could be abused to launch sophisticated cyberattacks. Since then, a string of incidents have surfaced in which US-made models broke free of their constraints and attacked companies and foreign government institutions, sparking weeks of debate over how model releases should be regulated.

What does Le Chonk actually do?

Mistral positions Le Chonk as a competitor to leading general-purpose models from the US and China, but it has tuned the model for specific domains: coding, cyberdefense, manufacturing, finance, and electrical engineering, among other niches. That specialization is deliberate, according to Guillaume Lample, Mistral's cofounder and chief scientist.

"There are a lot of areas where the other labs will not focus that much," Lample told WIRED. "There are so many domains in which you can improve models."

Mistral says Le Chonk is "very, very close" to some proprietary models. The company also claims it trained the model from scratch — a pointed contrast with Chinese labs, which the US government has accused of abusing distillation, the practice of training a smaller model on the outputs of a larger one, to close the performance gap with OpenAI and Anthropic.

Why does the open-weight claim matter for businesses?

Open-weight models carry a straightforward economic advantage: they cost only as much as the compute needed to run them. By narrowing the performance gap with leading proprietary models and offering a competitive alternative to Chinese releases, Mistral argues Le Chonk removes the last remaining reasons a business might hesitate to go open source.

Lample frames the stakes as broader than a US-versus-Europe contest. A business that relies on a proprietary model to repel cyberthreats, he argues, risks the sudden collapse of its defenses if access is cut off.

"Sometimes, people like to [make a big deal] over the US, versus Europe, versus China. But what really matters is to own the model — even for US companies," Lample said. "If you use a closed model, there is no guarantee it will still be there tomorrow."

Mistral wants to avoid being pigeonholed as a purely European champion. Its pitch is that fine-grained control over model access is a risk for any business, anywhere — including inside the United States.

How did Mistral get here?

The French lab has long operated with less capital and less compute than OpenAI and Anthropic, and it has generally trailed the US leaders in model performance, revenue, and release cadence. Its business model differs too: instead of charging a premium for closed, proprietary models, Mistral charges pay-as-you-go fees for running models through its cloud and deploys engineers to help customers tune models to their needs.

Recently, the gap has narrowed. In September, Mistral raised $3.3 billion at a $24 billion valuation — the largest funding round ever by a European tech company. Its earnings have reportedly increased 20-fold over the last year or so.

Lample dismissed any suggestion that Mistral has fallen out of the front rank. "Mistral is still in the race of getting the best model," he said. "This is the main message."

What role does geopolitics play?

Mistral's upswing coincides with growing friction between the US and its transatlantic allies — over tariff policy, Greenland, and the policing of American tech firms — alongside an increasingly fractious debate over who gets access to frontier AI.

The White House has reportedly asked American labs to withhold unreleased models even from the UK's AI Safety Institute, which had previously helped evaluate models for safety risks. That decision underscored for European policymakers that Washington can unilaterally revoke access to its most capable systems — an opening for a Europe-based lab that gives its weights away.

Andrea Renda, director of research at the Centre for European Policy Studies, told WIRED in July: "The continental strategy of the EU to become more technologically sovereign … and the increased hostility of the US is a magic formula that all of a sudden puts Mistral — whose performance has not been spectacular — in a favorable position."

Renda's assessment cuts both ways: Mistral's position owes as much to geopolitics as to raw capability. Le Chonk is the company's attempt to change that equation before the window shifts — proving it can match frontier performance while letting customers own the model outright.