White House orders Anthropic to shut down Fable 5 and Mythos 5 models over cybersecurity concerns

Anthropic removed access to its Fable 5 and Mythos 5 AI models for all users on June 12, 2026, following a government order to block foreign national access over national security concerns. The models had launched just three days earlier on June 9.

What triggered the shutdown

According to The Wall Street Journal, Amazon security researchers discovered methods to jailbreak Fable 5 to generate information potentially useful for cyberattacks. Amazon CEO Andy Jassy shared these findings with the White House, which then ordered Anthropic to block foreign access to both models.

A separate Semafor report claims the White House feared a China-linked group had accessed the models, though the White House has not confirmed this and the Trump administration's public statements focused solely on jailbreak concerns. Anthropic states the government "did not provide specific details of its national security concern" and that vulnerability evidence was only provided verbally.

Anthropic's response

The company opted to shut down access for all customers rather than implement foreign-access restrictions. In its statement, Anthropic said: "We are complying with the government's legal directive and are removing access to Fable 5 and Mythos 5 for all users. However, 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 export control order also blocked Anthropic's own foreign-born employees from accessing the models they developed.

About the models

At launch, Anthropic described Fable 5 as having "capabilities exceed those of any model we've ever made generally available," with particular strength in software engineering, knowledge work, and vision tasks. Claude Mythos 5 used the same underlying model but with "safeguards lifted in some areas."

Anthropic claims the discovered vulnerabilities were minor and exist in other models including GPT 5.5, though pricing, context window size, and benchmark scores for Fable 5 and Mythos 5 were not disclosed before the shutdown.

What this means

This marks the first time the U.S. government has forced an AI company to recall deployed models from commercial users. The shutdown demonstrates Washington's willingness to exercise emergency export controls over AI systems, even for models already in widespread use. For AI companies, the incident highlights regulatory risk from both technical vulnerabilities and geopolitical concerns. The broad order affecting Anthropic's own researchers also raises questions about how export controls will be implemented when much of the AI industry's workforce is foreign-born.