U.S. Lifts Export Controls on Anthropic's Fable 5 and Mythos 5

The U.S. Commerce Department withdrew export controls on Anthropic’s most capable models at the end of June, clearing the way for Fable 5 to return to users worldwide on July 1 and for Mythos 5 to be restored to a limited set of U.S. organizations with government approval.

Anthropic had disabled access to both models after the department issued an export-control directive on June 12, a move that required the company to cut service for many users just days after the models’ public debut.

The June 12 notice was an export-control action stemming from the Bureau of Industry and Security, which signaled that a license would be required for foreign persons to access the models while regulators evaluated potential misuse risks.

Anthropic said it began restoring Fable 5 on July 1 across its Claude platform — including Claude.ai and developer endpoints — after getting notice that the Commerce Department had lifted the controls. The company also said it had restored Mythos 5 access to a set of U.S. organizations following government approval.

According to Anthropic’s public post, Mythos 5 shares the same underlying architecture but is offered with fewer safety restrictions; the company said the narrower rollout for Mythos follows a government review completed June 26.

Commerce Secretary Howard Lutnick posted on X that his office “worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the U.S. Government and strengthen America’s leadership in AI,” language anthropic and media outlets reported.

U.S. officials initially flagged a technique that could be used to bypass safety restrictions — often called a “jailbreak” — and said the models posed potential cyber and misuse risks that required further analysis. That concern prompted the unusually swift enforcement action in mid-June.

In response, Anthropic implemented technical mitigations before redeploying Fable 5, including a classifier designed to block prompts that could identify software vulnerabilities or produce exploit code, and said it would coordinate safety testing with other major cloud and AI providers.

Policy analysts noted the episode is a test case for U.S. export-control authority over commercial AI models, raising questions for other labs and cloud providers about how regulators will assess risks and grant access going forward. Several think-tank and law-firm analyses highlighted how the June directive created legal and operational uncertainty for the industry.

Internationally, the Commerce Department’s reversal removes the immediate barrier that prevented non-U.S. users from using Fable 5, but it also underlines that model deployments can be paused or restricted while agencies evaluate national-security implications. Observers say the episode will shape how companies plan global rollouts.

Anthropic said its restoration plan includes rolling out updates to safety tooling, monitoring model behavior, and offering Mythos 5 only to vetted partners and government-approved organizations in the United States. The company emphasized cooperation with regulators as part of its path back to wide availability.

Industry coverage noted that other models of similar size and capability were able to reproduce some jailbreak-style demonstrations, a point used to argue that risks were not unique to Anthropic’s releases and that mitigation practices would need to be broader than a single company’s fixes.

For regulators, the case offers a blueprint: rapid technical review, a temporary restriction, and then phased restoration after mitigations and inter-agency consultation. Experts say turning that informal process into clearer rules would reduce surprise interruptions for developers and customers.

What to watch next is whether the Commerce Department or Congress moves to codify criteria for export controls on AI models, and how other labs will adapt release practices, safety testing, and transparency in the months ahead. For now, Fable 5 is back in global hands and Mythos 5 is available again under tighter U.S. oversight.