You should read the red team report: https://red.anthropic.com/2026/mythos-preview/

Anthropic's Frontier Red Team published a technical report (April 2026) detailing how their unreleased model, Claude Mythos Preview, autonomously identifies and exploits critical security vulnerabilities at unprecedented scale — discovering thousands of zero-days across every major operating system and web browser, including decades-old flaws, with over 83% exploit reproduction success on the first attempt. Due to its dual-use danger, Anthropic is withholding the model from public release and instead deploying it under controlled access through Project Glasswing, a defensive cybersecurity initiative with roughly 40 partner organizations including AWS, Apple, Google, and Microsoft. The red team report serves as an industry warning that similar capabilities are expected to proliferate across AI providers within 6–18 months, making coordinated defensive action urgent.

Related

• Again, if you care about computer security, read the red team report: https://red.anthropic.com/2026/mythos-preview/
• Curious how many large organization CISO offices have taken the Mythos red team reports as the red alert that it is. (I suspect very few) Ba…
• this is interesting. 1. Did Anthropic forget to run a control? 2. Where does this leave us?
• link to @HeidyKhlaaf’s sharp analysis:
• Folks, you can relax. Mythos is not some off-trend exponential gain. And the gains weren’t about recursive self-improvement. Good thread fro…
• My considered opinion is that The Mythos stuff was mostly a myth. Take it as a serious warning sign that we need to get our act together wit…

Source: safety