Anthropic Restricts Claude Mythos Over Zero-Day Finds

Anthropic announced the restricted release of its Claude Mythos Preview model on Monday. The AI system can autonomously identify and exploit thousands of zero-day vulnerabilities in major operating systems ^[1][2]. These critical security flaws were previously unknown to developers ^[1]. Safety concerns prompted Anthropic to limit access to a defensive consortium known as Project Glasswing ^[1]. This group includes major tech partners like Amazon, Google, and Microsoft ^[1][3]. The partners will use the tool to patch systems and improve global cybersecurity defenses ^[3]. Experts believe Mythos represents a major leap in autonomous reasoning and technical exploitation ^[2]. Anthropic stated that restricting the model is necessary to prevent potential misuse in large-scale cyberattacks ^[2][3]. The model remains unavailable to the general public ^[3].