If you've been following the AI space lately, you've probably seen the headlines. And yes, it's as wild as it sounds!
On June 12, 2026, the U.S. government forced Anthropic to pull its two newest AI models offline. Claude Fable 5 and Claude Mythos 5 had launched just three days earlier... and then, just like that, they were gone.
Here's what happened.
The Models
Mythos 5 is Anthropic's most powerful "frontier" model. When first announced in April, the company said it was too capable at hacking to release immediately. Fable 5 is essentially the same underlying model, but with added safeguards designed to prevent it from being used for cybersecurity exploits.
The Shutdown
On June 12, U.S. Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei ordering the company to suspend access to both models and block foreign nationals from using them, whether they live inside or outside the U.S., Anthropic determined it had no way to filter foreign nationals from U.S. users in real time, so it disabled both models globally for every user.
The Reason (Sort Of)
The U.S. government hasn't publicly stated the full reason, but Anthropic says it believes the government became aware of a "jailbreak" – a method for bypassing the safeguards in Fable 5 that prevent its most powerful features from being used for cyberattacks. Anthropic disputes the government's framing, calling the jailbreak narrow, non-universal, and no worse than what competing models already do.
The Bigger Picture
This didn't come out of nowhere. Earlier in 2026, the Pentagon demanded Anthropic allow Claude to be used for all "lawful use cases", which could include autonomous lethal weapons and mass domestic surveillance. Anthropic refused. In response, the Department of Defense designated Anthropic a "supply-chain risk," effectively blacklisting it from military and intelligence contracts and triggering multiple lawsuits.
Amazon, which has invested billions in Anthropic and runs Claude through AWS, reportedly raised security concerns with the government before the crackdown, adding another layer of complexity, given that Amazon is both a rival and a major investor.
Where Things Stand
As of June 23, Fable 5 is still unavailable — eleven days after the shutdown. The President said on June 19 that Anthropic is no longer considered a national security threat, but the order has not been lifted. Negotiations between Anthropic and the Trump administration are reportedly ongoing.
Anthropic has said it believes the government should have the ability to block unsafe deployments, but only through a process that is transparent, fair, and grounded in technical facts. It argues this action doesn't meet that standard.
Whether Fable comes back soon or this drags on for months, one thing is clear: we just watched the U.S. government flip the off switch on a frontier AI model for the first time in history. The AI industry will never look quite the same.
