U.S. Government Orders Anthropic to Pull Two New AI Models Over National Security Fears

Anthropic has abruptly shut down access to its latest AI models, Fable 5 and Mythos 5, following a direct order from the U.S. government. The directive, issued Friday at 5:21 p.m. ET, requires the company to block all usage by any foreign national - including its own employees outside the U.S. - citing national security authorities. Anthropic complied by disabling the models for every customer, though it stressed that other offerings remain unaffected.

The timing is striking. Just days earlier, Anthropic had launched Fable 5 and Mythos 5 as its most advanced public models yet, boasting top scores across industry benchmarks. Fable 5 was particularly notable as the company's first release of such a powerful system to the general public, thanks to new guardrails that restrict outputs in high-risk areas. The models evolved from Claude Mythos Preview, a tool that drew intense interest from Wall Street and government officials in April for its cybersecurity capabilities. Anthropic had kept Mythos Preview limited to a small group of companies under a program called Project Glasswing.

Anthropic said the government offered no specifics on the national security threat. The company apologized for the disruption, adding in a statement: "We believe the government should have the ability to block unsafe deployments, as part of a transparent, fair process grounded in technical facts. This action does not meet that standard."

The move is the latest flashpoint between Anthropic and the U.S. government. Earlier this year, after negotiations with the Department of Defense collapsed, the DOD labeled Anthropic a supply chain risk - a designation typically reserved for foreign adversaries. That label forces defense contractors to certify they won't use Anthropic's Claude models in military work. Anthropic has sued the Trump administration to overturn the blacklisting, and that case remains active.

Source: CNBC