Mistral Open-Sources AI Model Despite Warning It Could Aid Cyberattacks
InfoWorld, Friday, October 30th, 2026
Mistral released the open weights of Large 4 despite its own warning that the model could provide uplift for real cyberattacks.
Mistral published the open weights of its trillion-parameter Large 4, nicknamed Le Chonk, weeks after the public preview.
The company said it saw the model refuse black-box attacks on live targets but sees potential for uplift in real attacks, and it is adjusting guardrails.
Safeguard updates will apply only to its managed API and future releases, not to downloaded copies, said VP of science Pierre Stock, who named no specific attack stages it helped with and said the biggest gains were defensive vulnerability work.
The article notes the UK AI Security Institute found open-weight models trailed closed ones on cyber tasks by four to seven months.