Anthropic backs Project Glasswing
Anthropic’s unreleased frontier model is reported to power Project Glasswing, a collaboration with AWS, Apple, Google, Microsoft, NVIDIA and CrowdStrike to hunt zero‑day vulnerabilities in critical software like operating systems and browsers. (x.com) The partnership frames a security use case for frontier models in proactive vulnerability discovery.
Anthropic says its unreleased Claude Mythos 2 Preview will power Project Glasswing, a coalition hunting zero‑day flaws in critical software. (anthropic.com) Launch partners named by Anthropic include Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, NVIDIA and Palo Alto Networks. (anthropic.com) Anthropic announced Project Glasswing on April 7, 2026 and said it has extended access to the model for more than 40 additional organizations that maintain critical infrastructure. (anthropic.com) The company says Mythos has already found “thousands of high‑severity vulnerabilities,” including issues in major operating systems and web browsers. (anthropic.com) Anthropic calls Claude Mythos 2 Preview a “frontier model” and says it can “surpass all but the most skilled humans” at finding and exploiting software vulnerabilities. (anthropic.com) To limit risk, Anthropic is gating access to the model and committing up to $100 million in model usage credits plus $4 million in direct donations to open‑source security organizations. (anthropic.com) Cybersecurity outlets and research groups including CyberScoop and the Cloud Security Alliance have flagged Project Glasswing as evidence that AI‑driven vulnerability discovery is now practical and could be abused if it proliferates. (cyberscoop.com) Anthropic says Project Glasswing partners will use Mythos Preview to scan, triage and patch critical codebases, and that the company will share findings to benefit the broader industry. (anthropic.com) “The work of defending the world’s cyber infrastructure might take years,” Anthropic wrote, framing Glasswing as an urgent, defensive program against rapidly advancing AI capabilities. (anthropic.com)
