Mistral Opens Weights of Its Trillion-Parameter Large 4

French AI company Mistral has released the weights of Mistral Large 4, a trillion-parameter open-source model nicknamed "Le Chonk." Businesses and governments can now download and run it themselves. The model was first introduced earlier this month. The release comes as open-weight developers such as Mistral and closed-model providers like OpenAI and Anthropic compete harder, with Mistral arguing that businesses and governments should not have to rely on outside AI companies.

The weights are available on Hugging Face and through Mistral's API. Mistral is pitching the release at cybersecurity, where organizations want more control over their data and systems. "Amid growing cyber threats, ML4 brings state-of-the-art sovereign cyber defense capabilities to enterprises and governments globally," the company said on Thursday.

Performance claims

Mistral says the model is comparable in performance to GLM-5.3, a Chinese open-weight rival previously considered one of the best open-weight models in the world, particularly for cyber work. It says ML4 uses 47% fewer tokens per task than GLM-5.3.

The company also claims the model outranks every other open model on agentic coding tasks in Artificial Analysis' Coding Agent Index. CEO Arthur Mensch said AI should "be a tool for empowerment, not dominance." Mistral says its model reaches "state-of-the-art performance" at a "fraction of the compute and capital committed by closed labs."

Why open weights, according to Mistral

Pierre Stock, Mistral's vice president of science, told TechCrunch that open weights give organizations more control over how they use AI to defend themselves. "Access that can't be cut off by definition because the model is open-weighted," he said.

Stock said the model could be used to identify, coordinate and respond in real time to major security incidents. Because it is open, companies can deploy it on-premises in their own systems.

He said AI is helping security professionals find bugs at a much higher rate than before. It is important, he continued, to give the same capability to everyone as soon as possible, because some malicious actors might try to use their own technology to attack.

Stock also addressed recent incidents involving rogue AI agents. He acknowledged the risks but cautioned against overreacting. "We don't want to fearmonger people. We think it's a problem that is solvable," he said.

Source: TechCrunch