Trending 
Anthropic has suspended access to two of its advanced artificial intelligence systems, Fable 5 and Mythos 5, following a US national security directive that the company says forced it to restrict their worldwide availability immediately.
In a post published Friday, the AI developer said it was instructed to block access after receiving a government directive at 5:21 pm (2121 GMT), reportedly issued by US Commerce Secretary Howard Lutnick, according to Axios.
The directive, Anthropic stated, prohibited all foreign nationals, including employees working within the company, from accessing Fable 5 and Mythos 5 due to national security concerns. The firm said the scope of the order left it with no alternative.
“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” it said.
Fable 5 had only been launched publicly three days earlier. It is a restricted version of Mythos 5, a more powerful model that has not been widely released due to concerns over its potential capabilities. Mythos 5 has been made available only to selected companies under controlled conditions.
According to Anthropic, the government did not specify the exact security concern. However, the company said its interpretation is that officials are worried about a potential “jailbreaking” method that could bypass safeguards and enable the model to assist with hacking-related activity.
The firm added that it had reviewed the alleged technique but found no evidence that Fable 5 introduces unique risks beyond what already exists in other publicly available AI systems. It also noted that internal safety evaluations had not identified a “universal jailbreak” capable of fully bypassing its safeguards.
“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, warning that such a standard could stall the entire AI industry. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”
The suspension comes amid heightened tension between Anthropic and US authorities. The company has previously been involved in disputes with the Trump administration over concerns that its systems could be used for mass surveillance or autonomous weapons, contributing to the Pentagon cutting certain contracts with the firm.
The latest move places renewed scrutiny on the governance of frontier AI systems, particularly as governments tighten controls over technologies considered dual-use.
