The AI industry was surprised after Anthropic announced that the US government had ordered it to suspend access to its advanced Fable 5 and Mythos 5 models. The directive applies to all foreign nationals, whether they are inside or outside the United States, including foreign employees working at Anthropic.
According to the company, the US government issued the directive on June 12, and Anthropic immediately disabled both models to comply with the order. However, other Anthropic AI models remain unaffected.
“The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national,” Anthropic said.
Anthropic Pushes Back Against Jailbreak Claims
Anthropic said the government did not provide detailed evidence explaining its national security concerns. However, the company says reports that Fable 5 could be ‘jailbroken’ to bypass some of its safeguards likely triggered the government’s decision.
The company reviewed the reported technique and argued that it only exposed a small number of previously known vulnerabilities.
“We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities,” the company stated, adding that similar capabilities are already available in other AI systems, including OpenAI’s GPT-5.5.
Anthropic also defended its security practices, noting that Fable 5 underwent thousands of hours of testing involving US government agencies, the UK AI Safety Institute, private security firms, and internal red teams. According to the company, no tester has discovered a universal jailbreak capable of broadly bypassing the model’s protections.
The company said it relies on a “defense in depth” strategy that combines safeguards, monitoring, and data retention to limit misuse. It also warned that recalling commercial AI systems over narrow jailbreak concerns could set a dangerous precedent.
Reddit Questions the Move
The announcement quickly sparked debate online. Many Reddit users questioned whether the action was truly about national security, while others viewed it as a standard export-control measure similar to restrictions on encryption and cybersecurity technologies.
The discussion intensified after reports claimed the Trump administration was extending export controls on advanced AI models for foreign governments, companies, and individuals. The move reportedly followed claims that another company had successfully jailbroken Mythos and comes amid increasing government scrutiny of advanced AI systems.
For now, Anthropic says it is complying with the directive but believes the decision is a misunderstanding and is working to restore access as soon as possible.