# The company is restricting access as part of a cybersecurity initiative called Project Glasswing.

*fintech · news · 2026-04-08 · Mashable India*

## Key points

- Anthropic's Claude Mythos Preview AI model is restricted to select partners under Project Glasswing.
- The model can identify critical and previously undetected software vulnerabilities, including a decades-old OpenBSD bug.
- Anthropic limited access after internal debate and recent leaks of the model's details in a public data cache.
- Cybersecurity stocks reacted to the announcement due to concerns about the model's potential misuse by hackers.
- Anthropic is collaborating with major tech and cybersecurity firms but has no immediate plans for public release.

**Companies:** Apple, Google, Microsoft, Nvidia, Amazon Web Services, CrowdStrike, Palo Alto Networks
**Countries:** United States

[Read the full story on Mashable India](https://in.mashable.com/tech/108150/why-anthropic-is-limiting-access-to-its-most-powerful-ai-claude-mythos)

---

Canonical: https://newsio.io/n/c2a4c75d-a783-450d-ae46-7e04107a8ae9/the-company-is-restricting-access-as-part-of-a-cybersecurity-initiative-called-p
Summarized by Newsio from Mashable India. https://newsio.io/how-it-works