Anthropic Regains Global Access to Claude AI After US Lifts Ban

Anthropic will restore worldwide access to its most advanced AI models after US authorities lifted export restrictions imposed three weeks ago on national security grounds.

The Commerce Department withdrew controls on Claude Fable 5 and Mythos 5 models, allowing the company to resume global distribution starting July 2, according to details first reported by France 24. Secretary of Commerce Howard Lutnick notified Anthropic that the company had "worked with the US government to address risks associated with the Covered Models," Politico reported.

Rapid reversal follows security fixes

The government abruptly blocked access to both models on June 12 after identifying vulnerabilities in safeguards designed to prevent misuse. The restriction affected users globally, limiting access to these cutting-edge systems for nearly three weeks.

Just four days before the full restoration, Anthropic received authorization to provide Mythos 5 access to a small group of American cybersecurity firms. The Commerce Department's latest letter indicated satisfaction that Anthropic had "taken steps in close coordination with the US government to address the risks" associated with both models.

Why it matters

The episode reveals how the US government is asserting direct control over the release of powerful AI systems, treating frontier models as strategic assets requiring federal oversight. This marks a significant shift in AI governance, with Washington now actively intervening in product launches by private companies rather than relying solely on voluntary commitments. The precedent could reshape how AI labs develop and deploy advanced models globally.

Broader pattern of government intervention

Anthropic isn't alone in facing federal scrutiny. OpenAI has similarly complied with government requests to restrict its GPT-5.6 model to approved partners only. CEO Sam Altman acknowledged on Friday that "this isn't quite the process that we think is optimal" while explaining the limited launch.

The interventions stem from a June 2 executive order directing the federal government to create a voluntary framework for companies to test and release frontier AI models in collaboration with federal agencies. White House Chief of Staff Susie Wiles expressed gratitude Tuesday for tech companies cooperating on "advanced model access and guardrail testing and security."

CIA Director John Ratcliffe defended the administration's approach during a speech at the AWS summit in Washington, comparing advanced AI capabilities to "digital nuclear weapons." The comparison signals how seriously national security officials view the potential risks of unrestricted access to the most powerful AI systems.

Anthropic did not respond to requests for comment on the restoration of access.

These details were first reported by France 24 with AFP.