Anthropic restores global access to its AI models after US government lifts export controls

Anthropic will begin restoring global access to its most powerful AI models, Fable 5 and Mythos 5, after the U.S. Commerce Department lifted export controls on the technology, the company announced Tuesday. The reversal ends emergency restrictions imposed earlier this month by the Trump administration over fears the models could be exploited to bypass cybersecurity measures.

"We've received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5," Anthropic posted on X. "We'll begin restoring access tomorrow."

The removal of the controls follows weeks of coordination between the San Francisco-based lab and the government. Secretary of Commerce Howard Lutnick wrote in a letter to the company that the administration had "withdrawn" its previous restrictions and was satisfied Anthropic had "taken steps in close coordination with the U.S. government to address the risks" associated with the two models, Politico reported.

Why the models were restricted

The Commerce Department forced Anthropic to cut off access to the two models on June 12 after vulnerabilities were discovered in safeguards designed to prevent misuse. The abrupt move reflected a broader push by the Trump administration to limit the release of advanced AI systems that researchers said could help attackers bypass cybersecurity defenses.

CIA Director John Ratcliffe compared the capabilities of the most advanced AI models to nuclear weapons during a speech at the AWS summit in Washington. "It would be…not misplaced to refer to their capabilities as akin to digital nuclear weapons," Ratcliffe said, offering a tacit defense of the administration's hard line on controlling frontier AI technology.

A broader framework for AI governance

The president issued an executive order on June 2 directing federal agencies to develop a voluntary "framework" for private companies to test and release frontier AI models in partnership with the government. Anthropic, in a blog post Tuesday, called for a standardized process to assess critical vulnerabilities and said it would work with Amazon, Microsoft, Google and others on the effort.

As these frameworks take shape, government officials can deepen their understanding of AI risk and governance through the AI Learning Path for Policy Makers. The learning path is designed to give policy professionals the knowledge to navigate technical assessments and regulatory coordination.

Rival AI lab OpenAI has also complied with Washington's requests, restricting the release of a new model called GPT-5.6 to a limited set of approved partners. CEO Sam Altman posted on X that "this isn't quite the process that we think is optimal," signaling tension between innovation speed and government oversight that will likely shape future policy debates.

Why this matters for government professionals

This episode shows how quickly national security agencies can intervene in the release of advanced technology, and it underscores the growing demand for officials who can bridge the gap between AI capabilities and policy levers. The back-and-forth over export controls - from emergency cutoff to restored access in less than two weeks - demonstrates that the government is now a direct gatekeeper for the most powerful AI systems. For professionals in defense, intelligence, and regulatory roles, staying current on how these models are assessed and governed is no longer optional. The collaboration between Anthropic and the Commerce Department offers a real-world case of how voluntary frameworks and cross-sector coordination are being built in real time.