After announcing that its Mythos family of AI models was too powerful for public release, Anthropic launched a version of the program called Fable 5. Days later, it was suspended.
So, what happened?
Anthropic said via a statement on its website that it was ordered to suspend foreign nationals from using Claude Fable 5, including foreign national Anthropic employees. This comes after the US raised security concerns just days following its public release, with even Anthropic itself saying the technology was extremely powerful.
“The letter did not provide specific details of its national security concern. Our understanding is that the government believes it has become aware of a method of bypassing, or “jailbreaking”, Fable 5, Anthropic said in its online statement.
Jailbreaking refers to bypassing software restrictions that are designed to protect cyber networks – a process that can enable threat actors to access sensitive information.
Claude Fable 5 and Claude Mythos 5 were only announced on 9th June 2026, just last week, with safeguards attached. Anthropic said the release came with risks as, without safeguards, Fable 5 could be misused in areas like cybersecurity and cause “serious damage”.
The Mythos AI family was first released to a small number of companies for previewing and testing system vulnerabilities. Anthropic said this was because the tool is so intelligent that it could be dangerous for its ability to hack or exploit digital systems. This decision created unease, particularly amid concerns that next-generation AI models will contribute to accelerating the scale and sophistication of cyberattacks.
For Fable 5, the Claude company said it had implemented strong safeguards to reduce the likelihood it would be misused.
“In fact, our safeguards are so strong that many users have complained that they are overly broad,” the company said, adding that it had also worked with the US government, the UK AISI and multiple private third-party organisations to safeguard Fable.
“These tests showed that Fable’s safeguards are substantially more effective than those of any previously deployed model,” Anthropic added. “No testers have yet been able to find a universal jailbreak.”
The company has said it would comply with the US government’s new legal directive and has removed access to users. However, it noted that it disagreed that “a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people”.
It added: “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers. We believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear and grounded in technical facts. This action does not adhere to those principles.”
This isn’t the first time the US government has criticised Anthropic. The company is currently embroiled in a lawsuit with the country, after Anthropic wanted guarantees that its AI would not be used for mass surveillance or fully autonomous weapons that can kill without human control.
Capacity reported in March 2026 that the Pentagon had pressured Anthropic to weaken its ethical rules or face serious consequences, according to reports.
Writing at the time on Truth Social, US President Donald Trump said: “The Left-wing nut jobs at Anthropic have made a DISASTROUS MISTAKE trying to STRONG-ARM the [Pentagon] and force them to obey their Terms of Service instead of our Constitution.”
Speaking at the time, Anthropic said: “No amount of intimidation or punishment from the [Pentagon] will change our position on mass domestic surveillance or fully autonomous weapons.”
It added: “We have tried in good faith to reach an agreement with the [Pentagon], making clear that we support all lawful uses of AI for national security aside from the two narrow exceptions above.”
Related stories
