# Anthropic CEO Dario Amodei says his company's latest AI model has found tens of thousands of software vulnerabilities.

*genai · news · 2026-05-05 · CNBC*

## Key points

- Anthropic's AI model Mythos has uncovered tens of thousands of previously unknown software vulnerabilities.
- Chinese AI models are estimated to be six to twelve months behind Anthropic's Mythos in capability.
- Mythos found nearly 300 vulnerabilities in Firefox, compared to 20 found by earlier Anthropic models.
- Most vulnerabilities identified by Mythos remain undisclosed because they are still unpatched and exploitable.
- Anthropic released 10 new AI agents tailored for financial services and integrated them into Microsoft Office.

**Companies:** JPMorgan Chase, Anthropic
**Countries:** United States, China

[Read the full story on CNBC](https://www.cnbc.com/2026/05/05/anthropic-ceo-cyber-moment-of-danger-mythos-vulnerabilities.html)

---

Canonical: https://newsio.io/n/7fd927b1-e370-494f-a154-0fcc22d58f28/anthropic-ceo-dario-amodei-says-his-companys-latest-ai-model-has-found-tens-of-t
Summarized by Newsio from CNBC. https://newsio.io/how-it-works