Microsoft, Apple, Google, JPMorgan Among 50 Institutions Granted Early Access to Powerful Claude Mythos AI – Holes Remain in Public Rollout

Anthropic is providing early access to its unreleased Claude Mythos AI model to a select group of major technology and cybersecurity firms as part of a new initiative called Project Glasswing. The company says the model excels at identifying software weaknesses and security flaws, and it is limiting access to prevent potential misuse by bad actors.

According to Anthropic, the Claude Mythos Preview is an advanced AI model with strong agentic coding and reasoning skills that could reshape the cybersecurity sector. The initiative allows participating organizations to use the model for defensive security work and share their findings with the wider industry.

The initial launch partners include Amazon Web Services, Apple, Google, Microsoft, and Nvidia. More than 40 additional organizations responsible for building or maintaining critical software infrastructure are also receiving access, including CrowdStrike and Palo Alto Networks.

Anthropic says it does not plan to make the Mythos Preview generally available at this time. Instead, the company aims to safely deploy Mythos-class models at scale in the future once new safeguards are in place.

The existence of the Mythos model was first revealed in March 2026 when Fortune reported that Anthropic was developing and testing an unreleased model described in internal documents as “by far the most powerful AI model” it had ever developed. A draft blog post inadvertently made public last month warned that Mythos is “currently far ahead of any other AI model in cyber capabilities” and said it “presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders.”

Anthropic CEO and co-founder Dario Amodei discussed the company’s approach during the 56th annual World Economic Forum meeting in Davos, Switzerland, on January 20, 2026. Anthropic’s head of research product management, Dianne Penn, told CNBC that the company viewed Project Glasswing as a first step to give cyber defenders a head start on emerging threats.

Cybersecurity stocks declined following the Fortune report, which highlighted both the model’s advanced capabilities and the significant risks it poses if misused. Anthropic says it formed Project Glasswing because it believes the Mythos Preview could significantly improve defensive security efforts across critical systems.