Anthropic Limits Claude Mythos to US Gov and 40+ Partners

Anthropic has decided to limit access of its advanced model, Claude Mythos, to only the US government and a circle of more than 40 Project Glasswing partners. Ollie Chang's analysis for DigiTimes Asia, published May 15, 2026, puts this decision at the center of a widening access fight.

The analysis states that the restriction has broad implications for global users and policymakers. Chang examines how the move fits into larger patterns around AI, cybersecurity and government involvement.

According to the analysis, the decision signals that leading AI systems are now being treated as strategic assets. It is reshaping who can compete, defend, and innovate worldwide.

The report is titled "Mythos sparks access fight as AI models become strategic assets." It appears alongside related coverage including Anthropic's launch of Project Glasswing to fend cyber threats with the new Mythos model and the company's ban on Chinese-controlled entities from accessing Claude models.

Further related stories cited in the source address China's cybersecurity AI charging ahead despite the US model lockout, US security agencies adopting Anthropic's Mythos despite a Pentagon risk label, and the Mythos AI model triggering global financial and regulatory scrutiny amid cybersecurity concerns. The piece is categorized under AI, IT, CE, software and big data, with tags including analysis, Anthropic, Claude, cybersecurity, geowatch and government.