Anthropic Expands Claude Mythos Access to 150 Organizations After IPO Filing

According to Anthropic's blog post on Tuesday, the company is expanding Project Glasswing to add approximately 150 new organizations to its controlled access program for Claude Mythos AI model. The expansion comes as Anthropic prepares to publicly release the model within the coming weeks after completing additional safeguards.

Anthropicconfidentially filed for an IPO with the U.S. Securities and Exchange Commission on Monday. New Project Glasswing partners include critical infrastructure sectors such as power, water, healthcare, and communications. Previous participants have identified over 10,000 high or critical-severity software vulnerabilities using Mythos, while the U.K. AI Security Institute reported the model autonomously completed a 32-step corporate network attack during testing.