Anthropic unveiled Claude Mythos Preview, a generative AI model claimed to discover vulnerabilities and autonomously develop exploits for software systems. The model is being released to a limited consortium (Project Glasswing) including Microsoft, Apple, Google, and the Linux Foundation. Industry reactions vary—some argue existing AI agents already enable similar capabilities, while security practitioners highlight Mythos Preview's specific strength in identifying exploit chains.
Safety
Anthropic’s Mythos Will Force a Cybersecurity Reckoning—Just Not the One You Think
Anthropic releases Claude Mythos Preview, an autonomous vulnerability discovery and exploit development tool, to a limited consortium including Microsoft, Apple, Google, and the Linux Foundation—sparking debate over weaponized AI security research.
Friday, April 10, 2026 12:00 PM UTC2 MIN READSOURCE: WIRED AIBY sys://pipeline
Tags
safety
/// RELATED
ModelsApr 7
Assessing Claude Mythos Preview's cybersecurity capabilities
Anthropic launches Claude Mythos Preview with zero-day discovery and exploitation capabilities, backed by Project Glasswing, a coordinated effort to secure open-source software.
ModelsApr 7
Anthropic's Project Glasswing - restricting Claude Mythos to security researchers - sounds necessary to me
Anthropic gates Claude Mythos—which discovers thousands of high-severity OS and browser vulnerabilities—to security researchers via Project Glasswing before broader release.