Solutions  ·  2026-04-12

Anthropic Launches Project Glasswing with Claude Mythos Preview for Defensive Vulnerability Hunting

SolutionsHigh impact
Anthropic launched Project Glasswing on April 7, providing restricted access to Claude Mythos Preview—a frontier model with exceptional vulnerability-finding capabilities—to a coalition of 40+ organisations including Apple, Amazon, Microsoft, Google, and major cybersecurity firms. Anthropic committed $100M in usage credits and $4M in donations to open-source security.
Mythos Preview has identified thousands of previously unknown zero-day vulnerabilities in every major OS and browser. This represents a paradigm shift where frontier AI models are deployed defensively to find and patch critical flaws before adversaries exploit them.
Critical infrastructure operators, major software vendors, open-source maintainers, and government cybersecurity teams. Organisations maintaining critical codebases should evaluate whether they qualify for Glasswing access.
Sources
Anthropic - Project GlasswingFortune - Anthropic Claude Mythos ModelTechCrunch - Anthropic Mythos AI Model Preview SecuritySimon Willison - Project Glasswing Commentary
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →