Claude Mythos Preview Released with Advanced Cybersecurity Capabilities Under Restricted Access

Anthropic released Claude Mythos Preview, its most advanced frontier model, demonstrating significant improvements in reasoning, coding, and cybersecurity capabilities. Due to its powerful ability to autonomously discover and exploit zero-day vulnerabilities in major operating systems and web browsers, the model is restricted to approximately 50 vetted organizations including Google, Microsoft, Amazon, and major financial institutions through Project Glasswing. Internal testing revealed Claude Mythos could complete a 32-step corporate cyberattack simulation that typically requires 20 hours for human experts, making it the first AI model to autonomously pass such comprehensive security assessments.