Anthropic disables Fable 5 and Mythos 5 after US order on foreign access

WASHINGTON: Anthropic said on Friday it would shut off access to its most advanced artificial intelligence models after the US government directed the company to suspend availability for foreign nationals, citing national security concerns.

In a statement, the company said it had received an export control directive ordering it to halt access to Fable 5 and Mythos 5 for all foreign nationals, but had not been given detailed information about the security concern behind the move. Anthropic said it understood that the government believed there was a way to bypass, or jailbreak, a safeguard designed to stop Fable 5 from being used to identify software vulnerabilities.

The company said the government had provided only verbal evidence of what it described as a limited and non-universal jailbreak. Anthropic disputed the decision to withdraw a commercial system already deployed widely, saying the issue identified did not justify such a broad action.

“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,” the company said.

Compliance order and industry implications

Anthropic said the immediate effect of the directive was that it had to disable the two models for all customers in order to comply, while access to its other models would continue unchanged. The company added that it believed the matter stemmed from a misunderstanding and said it was trying to restore access as quickly as possible.

“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,” Anthropic said.

The company also argued that if the same threshold were applied across the sector, frontier AI developers would effectively be unable to roll out new models.

“If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” the company said.

A US official confirmed that the Commerce Department had issued the export control directive requiring suspension of all access to Fable 5 and Mythos 5 by foreign nationals. Amazon’s cloud division AWS said later on Friday that Anthropic had asked it to revoke access to the models for all users in all regions.

Tensions with Washington

The development comes as signs had emerged that a dispute between Trump administration officials and Anthropic, which confidentially filed for a US initial public offering last month, was easing in parts of the government. The relationship had deteriorated this year after Anthropic refused to let the US military use its AI systems for domestic surveillance and fully autonomous weapons systems. In response, the government placed the company on a supply chain blacklist that is due to take effect later this year.

The move also represents a significant widening of US efforts to curb the AI capabilities of foreign adversaries. For years, American export controls have focused primarily on chips and other hardware underpinning AI, rather than access to AI models themselves.

The dispute underscores increasing friction between AI companies and regulators over how to judge the risks posed by jailbreaks. Anthropic had as recently as Wednesday called for stronger US oversight of AI, including powers to stop models seen as carrying unacceptable risks, but said the action taken on Friday did not meet standards of fair and evidence-based regulation.

Pentagon chief information officer Kirsten Davies wrote on X that the Defense Department backed putting national security first.

“Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always,” Davies said.

Fable 5 launch and cybersecurity concerns

Earlier this week, Anthropic introduced Claude Fable 5, which it described as part of a new capability tier called Mythos-class. The model includes safeguards restricting use in high-risk fields such as cybersecurity, which Anthropic said some users had criticised as overly broad.

Experts have said Mythos models could, if misused, sharply increase the pace of sophisticated cyberattacks, especially in industries such as banking that depend on complex and often ageing technology systems. Anthropic said it had worked with the US government and others on safety issues before launching Fable, and added that rival AI models had shown a similar ability to find minor coding bugs.

Dean Ball, a former White House official who helped develop the administration’s AI Action Plan issued in the summer of 2025, said in a post on X that the order appeared to mean all non-Americans, including those residing in the US, would be barred from using Anthropic’s newest models.

“This means you should expect to have to prove your citizenship to use Anthropic models,” Ball said.

Several prominent Anthropic figures, including co-founder Chris Olah, AI researcher Andrej Karpathy and philosopher Amanda Askell, were born outside the United States. Reuters said it could not determine their citizenship status, while an Anthropic spokesperson declined to say whether such employees would also lose access to the AI models.