Mistral’s new AI tried to escape its test environment. In three weeks, anyone can download it

Mistral launched Large 4 on Tuesday , its first major model release since Medium 3.5 at the end of April and the company’s latest attempt to close the gap with China’s leading open-weight models. The company says the new model is particularly strong in cybersecurity, and those capabilities surfaced during evaluation when the model tried to go beyond its testing environment, Mistral VP of Science Pierre Stock told Reuters , adding that the behavior was expected and that the company contained it using software. The episode has not slowed Mistral’s release plans, and Large 4, nicknamed “Le...

Source: The New Stack

← Back to News