Anthropic has disabled access to its most advanced AI models after the US government issued an export control directive suspending access to Fable 5 and Mythos 5 for foreign nationals.

In a statement, the AI company said the directive applied to any foreign national, “whether inside or outside the United States”, including foreign national Anthropic employees. Anthropic said the effect of the order meant it would have to remove access to the models for all users in order to ensure compliance.

“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, whether inside or outside the United States, including foreign national Anthropic employees,” the company said.

“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected.”

According to Anthropic, the directive was received on Friday, June 12. The company said the letter did not provide specific details of the national security concern, but that it understood the government believed it had become aware of a method of bypassing, or “jailbreaking”, Fable 5.

Anthropic said it reviewed a demonstration of the technique being used to identify a small number of previously known, minor vulnerabilities. The company said those vulnerabilities “all appear relatively simple”, and that other publicly available models were able to discover them without requiring a bypass.

The order follows the launch earlier in the week of Claude Fable 5, a model Anthropic described as part of a new capability tier called “Mythos-class”. The company said the model had been accompanied by safeguards restricting its use in higher-risk areas including cybersecurity, though it acknowledged that some users had complained the guardrails were “overly broad”.

Anthropic said it had worked with the US government, the UK AISI, multiple private third-party organisations and internal teams to red-team Fable’s safeguards for thousands of hours before launch.

Recommended reading

• Anthropic Investigates Claimed Unauthorised Access to Mythos
• Anthropic’s Mythos Lands It a White House Meeting
• Is Mythos as Dangerous as Anthropic Claims?

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

“No testers have yet been able to find a universal jailbreak—a jailbreak method that can very broadly bypass the model’s safeguards, unblocking a wide range of cyber capabilities.”

However, Anthropic also said it does not believe perfect jailbreak resistance is currently possible for any model provider. The company said safeguards used across the industry remain vulnerable to non-universal jailbreaks, which may elicit some cyber information in specific circumstances, and that universal jailbreaks may eventually be found.

Anthropic said it had therefore adopted a “defense in depth strategy” for Fable 5, aiming to make jailbreaks either narrow or expensive to produce, while combining this with monitoring designed to detect and shut down successful attacks.

The company said this was also why it had required 30-day retention of customer data with Fable, describing the policy as a change that carries “real costs” with customers but supports research and mitigation around jailbreaks.

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

Recommended reading

• Anthropic Investigates Claimed Unauthorised Access to Mythos
• Anthropic’s Mythos Lands It a White House Meeting
• Is Mythos as Dangerous as Anthropic Claims?

The company said it had not received disclosure of “a concerning non-universal potential jailbreak that led to a harmful result”, adding that the potential jailbreaks disclosed to it were either benign or minor findings that provided “no Mythos-specific uplift”.

“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.

“Our understanding is that one potential jailbreak was shared with the government. We have reviewed a report that we believe is the basis of the government’s directive and validated that the level of capability displayed there is widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe. We will share more details over the next 24 hours.”

The directive marks a significant escalation in US efforts to restrict access to advanced AI capabilities. Until now, US export controls have largely focused on the chips and tools used to develop AI systems, rather than restricting access to AI models themselves.