Anthropic Scales Claude Mythos Vulnerability Program to 150+ Organizations in 15+ Countries
Key Takeaways
- ▸Claude Mythos now powers vulnerability discovery for ~150 critical infrastructure organizations across 15+ allied nations, representing significant international security coordination
- ▸The program explicitly targets organizations serving as infrastructure for governments and billions of people—a security-first approach to responsible AI deployment in high-stakes domains
- ▸Expansion demonstrates rapid commercialization of frontier AI capabilities in specialized cybersecurity use cases, while Anthropic races to establish safeguards before competitors develop equally capable models
Summary
Anthropic is expanding Project Glasswing, its joint initiative to identify and fix critical software vulnerabilities using AI, to approximately 150 new organizations across more than 15 countries. The expansion leverages Claude Mythos, Anthropic's most powerful model, which can identify thousands of zero-day vulnerabilities. The initial 50-partner pilot launched in early April with select government and private-sector participants; the new cohort expands to critical infrastructure sectors including power, water, healthcare, communications, and hardware that were underrepresented initially.
The program targets organizations whose security failures would have catastrophic global consequences—partners estimate that successful attacks on their codebases could affect over 100 million people. New participants include multinational firms like Samsung, SK Hynix, and Okta, as well as strategic institutions like NATO and the EU's cybersecurity agency ENISA. Participating countries include the U.S., Canada, Australia, France, Germany, Japan, South Korea, India, and other U.S.-allied nations. The expansion comes as Anthropic filed confidentially for an IPO following a $65 billion funding round, and as rival OpenAI released its own cybersecurity-focused model, GPT-5.5-Cyber.
- Project Glasswing reflects a shift toward collaborative, government-aligned AI initiatives for national security rather than purely commercial product rollouts
