Anthropic has developed an AI model named Claude Mythos with advanced capabilities in identifying and exploiting security vulnerabilities. This model has discovered thousands of previously unknown flaws across major operating systems and software, including a 27-year-old vulnerability in FreeBSD and a 16-year-old flaw in FFmpeg. Anthropic has deemed the model too dangerous for public release due to its potential for harm and is currently sharing it with a select group of large tech and finance companies to assist in patching these discovered vulnerabilities. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
RANK_REASON Frontier-lab model release with system card detailing unprecedented security capabilities.
