# It stated the model doesn’t just write code better, it can also find and.

*genai · news · 2026-04-08 · Outlook Business*

## Key points

- Anthropic's Claude Mythos Preview model can autonomously find and exploit software vulnerabilities with high efficiency.
- The model has already discovered thousands of zero-day vulnerabilities, including decades-old bugs in major systems.
- Anthropic launched Project Glasswing, a coalition with major tech firms, to proactively secure critical software infrastructure.
- Anthropic will not release Claude Mythos Preview publicly but will use it to develop cybersecurity safeguards.
- Project Glasswing partners will receive up to $100 million in model usage credits and open-source security donations.

**Companies:** Amazon Web Services, Apple, Google, JPMorgan Chase, Microsoft, NVIDIA, Broadcom, Cisco, CrowdStrike, Palo Alto Networks
**Countries:** United States

[Read the full story on Outlook Business](https://www.outlookbusiness.com/deeptech/anthropics-claude-mythos-breakthrough-project-glasswing-launched-to-prevent-ai-cyber-crisis)

---

Canonical: https://newsio.io/n/925e89dc-1ebf-4a6a-b007-ca7abebdc708/it-stated-the-model-doesnt-just-write-code-better-it-can-also-find-and
Summarized by Newsio from Outlook Business. https://newsio.io/how-it-works