Mistral opens weights of trillion-parameter Mistral Large 4

Mistral first introduced Mistral Large 4 earlier this month. The weight release follows a broader argument by open-weight developers that businesses and governments should not have to rely on outside AI companies for advanced capabilities. In cybersecurity, that argument may carry more weight because organizations want control over their data and systems. CEO Arthur Mensch framed AI as "a tool for empowerment, not dominance," and Mistral argues it can achieve state-of-the-art performance at a fraction of the compute and capital committed by closed labs. Pierre Stock, Mistral's VP of Science, said open weights give organizations access to frontier intelligence to defend against cyberattacks that cannot be cut off by definition. He added that companies can deploy the model on-prem in their own ecosystem to identify, coordinate, and respond in real time to major security incidents. Stock also addressed recent incidents involving rogue AI agents, acknowledging the risks but cautioning against overreacting: "We don't want to fearmonger people. We think it's a problem that is solvable."