The White House is considering opening up Anthropic's most powerful model, Mythos, to U.S. federal agencies, as the Office of Management and Budget is developing a special security protocol.

According to Dynamic Beating monitoring, an internal memorandum reveals that the White House is preparing to open Anthropic's most powerful model, Mythos, to key U.S. federal agencies. Federal Chief Information Officer Gregory Barbaccia of the White House Office of Management and Budget (OMB) has confirmed in an email to cabinet officials that the OMB is establishing relevant safeguards to allow agencies to begin using this tightly controlled AI tool.

Mythos is Anthropic's highest-capacity model to date and has not been publicly released. Previously, due to its transformative capabilities in the cybersecurity field (such as vulnerability discovery and attack tool creation), it was internally deemed by the company as posing a potential national security risk. This move by the White House indicates that the U.S. government, even amid concerns about cyber risks, must ensure that federal agencies can be at the forefront of leveraging cutting-edge AI productivity.