Anthropic has announced that it will restore worldwide access to its most advanced AI models after recent changes in US export policy.
The company confirmed that restrictions placed on its top models—reportedly named Claude Fable 5 and Claude Mythos 5—have been lifted following a review by US authorities. As a result, access to these models will now be gradually reinstated for users globally.
What changed
According to reports and a post shared by Anthropic on X, the US Department of Commerce has withdrawn earlier export control restrictions that had temporarily blocked the use of these advanced models. The company stated it would begin restoring access soon and would provide further updates as rollout progresses.
The decision follows a policy shift in which officials reassessed the restrictions and allowed the models to be used without requiring special export licenses for internal or cross-border deployment.
Why the models were restricted earlier
The suspension came after US export control measures were introduced to limit access to highly advanced AI systems by non-US users. Anthropic reportedly faced challenges in verifying user nationality in real time, which led the company to disable access to the affected models entirely, including for some of its own international employees.
The move was part of broader tensions over how frontier AI systems should be regulated and distributed globally.
Resolution and restart
After discussions between Anthropic and US government officials, the restrictions were lifted. Reports indicate that both sides worked together to ensure compliance and address security concerns related to the models’ capabilities.
A senior US official also noted that coordination with Anthropic helped align policy goals while maintaining US leadership in artificial intelligence development.
What happens next
With the restrictions now removed, Anthropic is preparing to re-enable access to its advanced models in phases. The company says users should expect further updates as services return to normal operation.
The situation marks a return to global availability for its frontier AI systems after a brief period of limited access.
