Security
Anthropic's red team research demonstrates that frontier models (GLM-5.3, Claude Mythos Preview) have crossed a capability threshold in binary exploitation and cyberattacks—a meaningful governance and procurement signal for organizations evaluating model safety, procurement risk, and responsible deployment practices.
Read the full article at Simon Willison
CoFabrix summarises and comments on this story. The original reporting belongs to Simon Willison.
Find out where your organization stands -- and what to do about it.