Story perspectives

Anthropic Launches AI Cybersecurity Model with Major Tech Partners

4/7/2026

Summary

Story summary

On April 7, 2026, Anthropic announced Claude Mythos Preview, an AI model for cybersecurity under Project Glasswing.
The initiative aims to help companies identify software vulnerabilities while restricting access to prevent misuse.
Initial partners include Apple, Google, Microsoft, Nvidia, and Amazon Web Services, with more than 40 others involved.
The announcement cites concerns about capabilities and a stock decline; Anthropic is discussing applications with U.S. officials.

1 of 1

Story summary

On April 7, 2026, Anthropic announced Claude Mythos Preview, an AI model for cybersecurity under Project Glasswing.
The initiative aims to help companies identify software vulnerabilities while restricting access to prevent misuse.
Initial partners include Apple, Google, Microsoft, Nvidia, and Amazon Web Services, with more than 40 others involved.
The announcement cites concerns about capabilities and a stock decline; Anthropic is discussing applications with U.S. officials.