Anthropic announced it is taking its newly launched AI models, Claude Fable 5 and Mythos 5, offline following an export control directive from the US government issued Friday afternoon. The company said the order, citing national security concerns, required suspension of access to any foreign nationals, including Anthropic employees. To ensure compliance, Anthropic disabled access for all customers. The move represents a new escalation in tensions between the company and the Trump administration.

Earlier this year, the Department of Defense labeled Anthropic a “supply chain risk” after disputes over military use of its technology, effectively barring government agencies from using its products. Anthropic had recently released Claude Fable 5, a safeguarded version of its Mythos AI model designed to limit responses on cybersecurity, biology, and chemistry. The company said it received the government’s letter at 5:21 p.m. ET but that it contained no specific details about the security concern.

Anthropic stated that the government believes a method exists to “jailbreak” Fable 5, though the company described the vulnerability as narrow and minor. CEO Dario Amodei reiterated support for transparent government oversight but argued that this action did not meet those standards.