IT ProPortal

Anthropic reportedly withholds access to Mythos 5.1 from UK safety testing body

The decision marks the first time the AI Security Institute has been left out of pre-release evaluations of Anthropic models