US tells Anthropic to suspend Fable 5, Mythos 5 access for foreign nationals

The US government has directed Anthropic to suspend access to its Fable 5 and Mythos 5 artificial intelligence models for all foreign nationals, including Anthropic’s own foreign national employees, prompting the company to abruptly disable the models for all customers.

In a statement, Anthropic said the directive was issued under national security authorities and applied to foreign nationals “whether inside or outside the United States.” The company said access to its other models would not be affected.

“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” Anthropic said.

The move marks a rare and significant intervention by the US government in the deployment of frontier AI models, coming just days after Anthropic introduced its new Mythos-class systems. Fable 5 was positioned as the publicly accessible version, while Mythos 5 was subject to tighter controls.

Anthropic said it received the government directive at 5.21 pm ET on June 12, but added that the letter did not provide specific details of the national security concern.

“Our understanding is that the government believes it has become aware of a method of bypassing, or ‘jailbreaking’ Fable 5,” the company said.

According to Anthropic, it reviewed a demonstration of the technique being used to identify “a small number of previously known, minor vulnerabilities.”

 The company argued that these vulnerabilities were “relatively simple” and could also be found using other publicly available models.

Anthropic said it had put in place strong safeguards for Fable 5, particularly around cybersecurity-related misuse, and had worked with the US government, the UK AI Safety Institute, private third-party organisations and internal teams to red-team the model before launch.

“These tests showed that Fable’s safeguards are substantially more effective than those of any previously deployed model,” Anthropic said.

The company added that no tester had so far found what it described as a “universal jailbreak,” or a method that could broadly bypass the model’s safeguards and unlock a wide range of cyber capabilities.

“We suspect that perfect jailbreak resistance is not currently possible for any model provider,” Anthropic said. “Every safeguard used in the industry is vulnerable to non-universal jailbreaks.”

Anthropic said its strategy with Fable 5 was based on “defense in depth,” aimed at making jailbreaks either narrow or expensive to produce, while combining safeguards with monitoring to detect and shut down attacks. The company said this was also why it required 30-day retention of customer data for Fable, despite the cost of such a policy with customers.

“We stand by this defense in depth strategy. It reduces the risks posed by Fable, making them comparable to the risks of existing models already deployed across the industry,” Anthropic said.

The company pushed back against the government’s reasoning, saying it had not received evidence of a serious jailbreak that led to harmful results.

“To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws,” Anthropic said.

Anthropic said it had reviewed a report that it believed formed the basis of the government’s directive and found that the demonstrated capability was already widely available from other models, including OpenAI’s GPT-5.5.

The company said it was complying with the legal directive but disagreed with the decision.

“We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” Anthropic said. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”a

Anthropic also said it supports a government role in blocking unsafe AI deployments, but argued that such action should follow a statutory process that is “transparent, fair, clear, and grounded in technical facts.”

“This action does not adhere to those principles,” the company said.

The suspension could disrupt customers who had begun integrating Fable 5 and Mythos 5 into workflows, particularly in software development, research and enterprise automation. Anthropic said it was working to restore access.