Fable 5, Anthropic’s most powerful publicly available AI model, was pulled offline on Friday, 12 June 2026, after the United States government issued an export control directive barring all foreign nationals from accessing it.
The directive landed on Friday afternoon and forced one of the biggest AI companies on the planet to yank a flagship product within hours.
According to Anthropic, the order required it to block every foreign national from using Fable 5 and its sister model Mythos 5, no matter where in the world they sat.
What the US government ordered for Fable 5
The instruction came from the US Commerce Department, which leaned on national security authorities to justify it.
The letter reportedly arrived at 17:21 Eastern Time (23:21 SAST), and it did not spell out the specific national security worry that triggered such a drastic step.
The scope is what makes the order unusual. It does not just bar users in rival states. It covers any foreign national anywhere, including Anthropic staff who are not US citizens, which left the company with no clean way to comply other than switching the models off entirely.
Why Anthropic disabled Fable 5 for everyone
Because the platform could not isolate foreign nationals on the fly, Anthropic took the blunt route and shut everything down.
Anthropic said the models had to go dark “for all our customers to ensure compliance,” while making clear that every other Claude model stayed online.
The company did not hide its frustration. Anthropic apologised to customers for the disruption, called the whole thing a misunderstanding, and said it was scrambling to restore access as fast as possible.
It also flagged that the demand to police foreign nationals was effectively impossible to action surgically.
The jailbreak claim behind the Fable 5 ban
So what actually spooked Washington?
Anthropic’s read is that officials believe they found a way to jailbreak Fable 5, meaning a trick to slip past the guardrails baked into the model. The company says it reviewed a demonstration of the technique and was distinctly unimpressed by what it saw.
By Anthropic’s account, the so-called exploit only surfaced a handful of minor vulnerabilities that were already known, and the underlying capability is nothing exotic.
It noted that rival systems, including OpenAI’s GPT-5.5, can do the same thing, and that security defenders rely on these abilities every single day.
The kicker is how thin the evidence looks from where Anthropic is standing. The company says the government has so far offered only verbal evidence of the flaw, which boils down to asking the model to read a codebase and patch software bugs.
That is ordinary work for any coder.
Why the Fable 5 ban is a first
This is bigger than one product going dark for a weekend. It is, as far as anyone can tell, the first time a leading AI company has yanked a publicly deployed model offline because of a direct order from the federal government.
That is a precedent the whole industry will be watching.
Fable 5 was a milestone for Anthropic, the first time it released a model this advanced to the general public, sitting on top of the more restricted Mythos system.
Mythos has already turned heads in Washington and on Wall Street for its cybersecurity muscle, which is exactly the capability now under scrutiny.
What happens next for Fable 5
For now, Fable 5 and Mythos 5 stay switched off while Anthropic argues its case to the government and pushes to get them back online.
The company is betting it can convince officials the concern is overblown, but there is no public timeline for when, or whether, full access returns.
