French AI company Mistral has released the weights for its trillion-parameter open source AI model Mistral Large 4 (ML4), nicknamed “Le Chonk.” The release lets businesses and governments download and deploy the system themselves, and it strengthens Mistral’s argument that organizations shouldn’t have to rely on outside AI companies for advanced capabilities—especially in cybersecurity, where control over data and systems matters. As CEO Arthur Mensch put it, AI should “be a tool for empowerment, not dominance.”

Mistral says ML4’s performance is comparable to Chinese open-weight competitor GLM-5.3, previously considered one of the best open-weight models in the world for cyber work, while using 47 percent fewer tokens per task. The company also claims the model outranks all other open models on agentic coding tasks in Artificial Analysis’ Coding Agent Index. Mistral says it achieves this at a fraction of the compute and capital committed by closed labs. The weights are available on Hugging Face and via Mistral’s API.

Pierre Stock, Mistral’s Vice President of Science, said the open weights mean access that can’t be cut off, since organizations can deploy the model on-prem in their own ecosystem and use it to identify, coordinate, and respond in real time to major security incidents. He argued for giving the capability to everyone as soon as possible to strengthen the ecosystem, because malicious actors might leverage their own technology to attack. Addressing incidents involving rogue AI agents, Stock acknowledged the risks but cautioned against fearmongering, calling it a solvable problem—even though, as he said, “It’s a hard problem to have good security.”