Federal Agencies Seek Access to Anthropic’s Claude Mythos Preview

Federal agencies have requested access to Claude Mythos Preview, a powerful new AI model from Anthropic that can rapidly identify — and potentially create — new cyberthreats, according to reporting by The New York Times.

The request comes as the White House works to authorize a modified version of the model for use by major US federal agencies under safeguards designed to limit misuse, amid ongoing concerns about the model’s dual-use capabilities in cybersecurity.

Anthropic announced Claude Mythos Preview on April 7 under Project Glasswing, a controlled-access program for select technology and financial organizations, as confirmed by multiple sources including the company’s own disclosures and federal oversight memos.

Federal Chief Information Officer Gregory Barbaccia at the White House Office of Management and Budget (OMB) informed Cabinet department officials that the OMB is establishing protections to allow federal agencies to begin using the model, according to an internal memo reported by Bloomberg.

The memo did not commit specific agencies to deployment or provide a timeline, but emphasized collaboration with model providers, industry partners and the intelligence community to ensure appropriate guardrails are in place before any release.

The move occurs while the Department of Defense’s supply-chain risk designation against Anthropic, issued on March 3, remains in force. The D.C. Circuit refused to stay the designation on April 8, keeping the company barred from defense contracts even as civilian agencies are being positioned for access.