Assessing Claude Mythos Preview’s cybersecurity capabilities

Anthropic has announced the release of Claude Mythos Preview, a new general-purpose language model exhibiting advanced computer security capabilities. According to the company, the model performs strongly across the board and can identify and exploit zero-day vulnerabilities in every major operating system and web browser when directed by a user. Anthropic noted that over 99 percent of the vulnerabilities discovered by the model remain unpatched. The company launched Project Glasswing to share Mythos Preview with a limited group of critical industry partners and open-source developers to help secure important systems before such capabilities are broadly available. Testing showed that Mythos Preview represents a substantial leap in cybersecurity capabilities compared to earlier models like Opus 4.6, which had near-zero success rates in autonomous exploit development. Anthropic emphasized that these capabilities emerged naturally from general improvements in code, reasoning, and autonomy rather than explicit training. While the transitional period may present challenges, Anthropic believes that powerful language models will ultimately benefit defenders more than attackers once the security landscape reaches a new equilibrium.