Anthropic Blocks Mythos AI Release Amid Global Cybersecurity Alarm

On April 7, 2026, Anthropic announced it would withhold its latest frontier AI model, Claude Mythos, from general public release. The company cited catastrophic risks to national security and public safety, as the model possesses an autonomous ability to identify and exploit high-severity zero-day vulnerabilities across all major operating systems and web browsers, including a 27-year-old flaw in OpenBSD. To manage these risks, the company established Project Glasswing, a defensive consortium providing controlled access to over 40 organizations, including Microsoft, Google, Amazon, Apple, and JPMorgan Chase. Anthropic supported the initiative with $100 million in usage credits and $4 million in donations to open-source security groups like the Linux Foundation and Apache Software Foundation.

The rollout triggered urgent government responses. U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened emergency meetings with major Wall Street bank CEOs to warn of potential financial system destabilization. Similar coordinated efforts occurred in Canada and the UK. However, the company remains in a legal battle with the U.S. Department of Defense, which designated Anthropic as a supply-chain risk after the firm refused to allow its technology to be used for autonomous weapons or mass surveillance.

Industry reactions were polarized. While partners like CrowdStrike and Cisco praised the tool for accelerating defense, critics including OpenAI CEO Sam Altman and investor David Sacks characterized the restricted release as a fear-based marketing strategy to create artificial scarcity. Meanwhile, independent testing by the U.K. AI Security Institute confirmed that Mythos could execute complex, multi-stage simulated cyberattacks that typically require days of human effort, completing them in minutes.