The U.S. Commerce Department lifted export controls on Anthropic's Fable and Mythos artificial intelligence models on Tuesday, June 30 [1].

This decision restores international access to some of the company's most powerful AI tools. The move signals a shift in how the U.S. government balances the rapid global deployment of advanced AI with national security safeguards.

The restrictions had been in place since the models were originally suspended on June 12 [2]. According to some reports, these specific versions are known as Fable 5 and Mythos 5 [2]. The models were subjected to a review process to determine if their capabilities created vulnerabilities or risks to the state.

"The Commerce Department has determined that the Fable and Mythos models no longer present an undue national‑security risk," a U.S. Commerce Department spokesperson said [3].

Government officials indicated that the decision followed a rigorous assessment of the technology. A U.S. Commerce Department official said that lifting the curbs reflects confidence that the models meet export-control requirements [4].

Anthropic is now preparing to reopen the systems to global users. "We'll begin restoring access tomorrow," an Anthropic spokesperson said [5].

The timeline for the lift has seen slight variations in reporting, with some sources citing June 30 [1] and others noting the announcement on July 1 [4]. However, the consensus remains that the models are now cleared for international distribution after a brief period of suspension.

“The Commerce Department has determined that the Fable and Mythos models no longer present an undue national‑security risk.”

The brief suspension and subsequent release of Fable and Mythos demonstrate the U.S. government's intent to implement a 'review and release' mechanism for frontier AI models. By temporarily halting exports to conduct security audits, the Commerce Department is establishing a precedent for how it will manage the tension between American commercial competitiveness and the prevention of AI-driven security threats.