US Government Orders Anthropic to Suspend Access to Fable 5 and Mythos 5 Models

The United States government has issued an immediate export control directive to Anthropic, ordering the suspension of access to its Fable 5 and Mythos 5 artificial intelligence models for all foreign nationals. The order, which applies to individuals regardless of their location or employment status with the company, was received by Anthropic on Friday evening at 5:21pm ET. In response, the company abruptly disabled access to these specific models for all customers to ensure compliance with the federal mandate.

Anthropic stated that it disagreed with the necessity of the recall, noting that the government’s concern centred on a method of bypassing, or “jailbreaking,” Fable 5 to identify minor vulnerabilities. The company argued that these vulnerabilities are relatively simple and that other publicly available models, including OpenAI’s GPT-5.5, can discover similar issues without requiring such bypasses. Despite this assessment, Anthropic confirmed it was complying with the legal directive and removing access to Fable 5 and Mythos 5 for all users, while access to other Anthropic models remains unaffected.

The timing of the directive has drawn scrutiny regarding potential market manipulation. The announcement was made after futures markets had closed, a pattern noted by research firm The Kobeissi Letter, which has catalogued previous major geopolitical and trade announcements by the Trump administration that landed on Friday evenings or early Saturdays. This strategy is often employed to allow markets time to settle over the weekend before the full impact is reflected in trading hours.

Tensions between the Trump administration and Anthropic have been documented previously, including a declaration of the company as a supply chain risk that banned its use in government systems. This occurred despite Anthropic’s efforts to integrate with the Department of Defense and the continued use of its models in classified military operations. Critics suggest the current directive may be influenced by these political frictions, particularly as Anthropic prepares for an initial public offering, with competitors like OpenAI enjoying closer ties to administration figures.

The move has sparked broader debate about the precedent of governments restricting access to advanced artificial intelligence. Industry observers warn that if the strongest large language models are deemed too dangerous for general use, it could signal a shift where powerful AI becomes restricted to nation-state control. This raises concerns about the long-term viability of the AI investment boom, as the potential for unilateral government restrictions could undermine the projected value of data centre infrastructure and commercial AI development.