Mistral Opens Up Trillion-Parameter Model Weights, Aiming at Cyber Defense

French AI developer Mistral has published the weights for its trillion-parameter open source model, Mistral Large 4 (ML4), letting businesses and governments download and run the system on their own terms. Nicknamed "Le Chonk," the model was unveiled earlier this month.

The move sharpens competition with closed-model providers such as OpenAI and Anthropic. Mistral's argument is that organizations shouldn't depend on outside AI companies for advanced capabilities - a case that carries particular weight in cybersecurity, where control over data and systems matters most.

CEO Arthur Mensch framed AI as "a tool for empowerment, not dominance." Mistral says ML4 reaches state-of-the-art performance at a fraction of the compute and capital committed by closed labs.

Per the company, the model performs comparably to Chinese open-weight competitor GLM-5.3, previously regarded as among the best open-weight models, especially for cyber work - while using 47 percent fewer tokens per task, making it considerably cheaper to run. Mistral also claims ML4 outranks all other open models on Artificial Analysis' Coding Agent Index for agentic coding tasks.

The weights are available on Hugging Face and through Mistral's API. "Amid growing cyber threats, ML4 brings state-of-the-art sovereign cyber defense capabilities to enterprises and governments globally," the company said Thursday, adding that institutions can now "download, deploy, and own frontier AI to detect, analyze, and respond to threats across code and IT infrastructure."

Pierre Stock, Mistral's Vice President of Science, told TechCrunch that open weights give organizations access to frontier intelligence for defending against cyberattacks - access that "can't be cut off by definition because the model is open-weighted." He said the model could help identify, coordinate, and respond in real time to major security incidents, and that open sourcing lets companies deploy it on-prem in their own ecosystems.

Stock argued that AI-driven security capabilities should reach everyone as soon as possible to strengthen the ecosystem, since "some malicious actors might try to actually leverage their own technology to attack." Asked about 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. It's a hard problem to have good security."

Source: TechCrunch