UK AI Safety Institute review of Claude Mythos: able to autonomously complete a 32-step enterprise network attack simulation
The UK AI Safety Institute (AISI) latest evaluation shows that Anthropic’s Claude Mythos Preview can autonomously complete a full 32-step enterprise network attack simulation in a controlled environment. In expert-level CTF challenges, it achieves a 73% success rate, marking a key threshold being crossed in AI cyberattack capabilities.
(Background: Claude officially supports modifying Word files and saving workflows as skills, with the full integration of Microsoft Office’s three-piece suite completed.)
(Additional context: Anthropic AI Economic Index—tens of thousands of words report: the frequency of automated trading workflow execution has doubled; Claude is evolving from a tool into a life assistant.)
Table of contents
Toggle
CTF Evaluation: 73% expert-level attainment rate
動區BlockTempo·6h ago