Mistral unveils Le Chonk, a 1-trillion-parameter open-weight cybersecurity model
French AI lab Mistral just introduced Le Chonk (officially Mistral Large 4 or ML4), a massive 1-trillion-parameter AI model built to help fight cyber threats.
Unlike many closed-off models, Le Chonk is open-weight, so users get more control, something that's becoming more important as worries about AI security grow.
Le Chonk competes with GPT-6 Astra
Le Chonk is in public preview now, with its model weights releasing on October 27.
It was trained from scratch on 4,000 NVIDIA Grace Blackwell GPUs over two months, deployed in Mistral's own data centers in Europe, and can handle everything from finance to complex multimedia tasks.
Mistral is also making sure organizations can run the model wherever they need, including under strict EU data rules, giving users real flexibility and ownership over their cybersecurity tools.
This puts Le Chonk up against big names like GPT-6 Astra, offering a fresh, user-friendly option for staying safe online.