Anthropic scales Claude Mythos to 150 orgs, 15 countries for critical infrastructure

Anthropic is scaling Project Glasswing, its joint industry initiative to find and fix critical software vulnerabilities using AI, to approximately 150 new organizations across more than 15 countries. The initiative centers on Claude Mythos, which Anthropic calls its most powerful model yet, capable of identifying thousands of zero-day vulnerabilities over several weeks. Initially given to 50 partners including the U.S. government in April, the expanded cohort now includes organizations in power, water, healthcare, communications, and hardware—industries previously underrepresented. These organizations maintain codebases that other governments and businesses rely on, and Anthropic estimates a major attack could affect over 100 million people.

Specific organizations receiving access include U.S.-based Okta, South Korean companies Samsung, SK Hynix, and SK Telecom, as well as NATO and the EU’s ENISA. The expansion covers countries friendly to the U.S., such as Australia, Canada, France, Germany, and Japan. Anthropic filed confidentially for an IPO following a $65 billion funding round at a nearly $1 trillion valuation. The company is racing to establish safeguards, expecting rivals like OpenAI (with its GPT-5.5-Cyber) to soon match Mythos's capabilities.