An advanced Anthropic AI model named Mythos has reportedly identified critical weaknesses in “sensitive and secure U.S. government computer systems” during a testing exercise, according to an anonymous U.S. official.
The official stated that the tool broke into almost all of their classified systems within hours — not weeks.
Project Glasswing, which was announced on April 7, is described as an urgent attempt to apply the capabilities of the frontier model Mythos Preview for defensive purposes. However, the identification of vulnerabilities does not mean they were exploited.
