Anthropic expands Claude Mythos access to 150 orgs ahead of IPO filing

Editorial illustration for: Anthropic expands Claude Mythos access to 150 orgs as AI giant files for IPO

In brief

  • Anthropic adds ~150 organizations to Project Glasswing for Claude Mythos access and testing
  • IPO filing: Anthropic confidentially submitted SEC filing on Monday
  • Previous Glasswing partners identified 10,000+ critical vulnerabilities in Mythos Preview
  • Mythos-class models expected to release to customers within weeks after safeguards complete
  • Anthropic warns competitors may release similar models in 6-12 months without safety measures

Controlled access to powerful AI

Anthropic said it is adding roughly 150 organizations to its Project Glasswing program in a blog post on Tuesday. The new cohort includes organizations from critical infrastructure sectors, including power, water, healthcare, communications, and hardware. Each organization must meet Anthropic's security requirements before gaining access to the model.

The company emphasized the stakes of this work. "What each partner has in common is that a successful attack on their codebase could be catastrophic," Anthropic said in its announcement. "For most partners, we estimate that a major attack could affect more than 100 million people, with important ramifications for both global and national security."

Real-world impact so far

Previous Glasswing participants have already demonstrated the model's value. Organizations with earlier Mythos Preview access have identified more than 10,000 high or critical-severity software vulnerabilities. Mozilla reported that the model identified 271 vulnerabilities it fixed in the Firefox web browser. The U.K. AI Security Institute reported that Mythos autonomously completed a 32-step simulated corporate network attack during testing.

Claude Mythos emerged publicly in March after draft Anthropic materials leaked online. Anthropic expects to bring Mythos-class models to customers in the coming weeks once additional safeguards are completed.

The competitive race ahead

Anthropic flagged a looming challenge. The company said it expects other developers to release Mythos-class models within six to 12 months, potentially without comparable safeguards. That timeline underscores why Anthropic is pushing the responsible disclosure process now—before competitors flood the market with less carefully vetted versions.

Market sentiment reflects the uncertainty. Users on Myriad, a prediction market platform, estimate a 26.5% chance that Claude Mythos will be released to the public by the end of June, down from a peak of 58.5% the previous week. The drop suggests traders are pricing in a longer timeline for public availability.