Anthropic2026-04-16
Claude Mythos Cybersecurity Model Discovers Thousands of Critical Vulnerabilities Under Restricted Access

AI Analysis
Anthropic's Claude Mythos model, deployed under controlled 'Project Glasswing' access, has discovered thousands of major vulnerabilities in operating systems, web browsers, and other software. The restricted-access model represents a significant cybersecurity advancement. Separately, investors are flooding Anthropic with funding offers valuing the company at up to $800 billion.