Claude Mythos: Anthropic's AI Found Zero-Day Flaws in Every Major OS — What It Means for Your Security Career

# Claude Mythos: Anthropic's AI Found Zero-Day Flaws in Every Major OS — What It Means for Your Security Career

> **Quick answer:** Anthropic's new AI model Claude Mythos Preview autonomously found thousands of zero-day vulnerabilities in every major operating system and browser as part of Project Glasswing, including a bug that survived 23 years and 5 million automated security tests. Over 99% of these vulnerabilities are still unpatched. The model is not publicly available, but its existence fundamentally changes what security careers look like starting now.

Anthropic's Claude Mythos cybersecurity capabilities have done something the entire security industry has been building toward for decades: an AI that can autonomously find critical, previously unknown flaws in every major operating system, every major browser, and even in code written in supposedly "safe" languages like Rust. If you work in security, or are planning to, the timeline just accelerated.

## What Claude Mythos Found — And How Alarming the Numbers Are

Project Glasswing is Anthropic's restricted cybersecurity initiative, announced April 7, 2026. The centerpiece is Claude Mythos Preview, a model that was never released to the public and is only accessible to 40+ vetted organizational partners including AWS, Apple, Cisco, CrowdStrike, Google, JPMorganChase, Microsoft, NVIDIA, and Palo Alto Networks.

What the model found in Anthropic's own red team testing is genuinely staggering: