The US has allowed Anthropic to provide its Claude Mythos 5 AI model to more than 100 approved companies and institutions after a review of national security safeguards.
The government had earlier suspended access to the model over fears it could be used by military or intelligence entities in countries such as China and Russia. Restrictions continue for organisations outside the approved list.
The move reflects Washington's evolving approach to AI export controls—protecting sensitive technologies while allowing trusted commercial partners to access advanced AI models.
US Relaxes Export Restrictions on Anthropic's AI Amid Security Safeguards
The US has partially lifted export restrictions on Anthropic's flagship AI model, Claude Mythos 5, allowing access for approved companies after the startup addressed national security concerns raised by Washington
Summary
Advertisement
The US government has partially eased restrictions on Anthropic's most advanced artificial intelligence model, allowing the company to provide access to selected "trusted partners" just two weeks after imposing export controls over national security concerns.
According to a Commerce Department letter seen by Reuters, more than 100 companies and institutions, including several Fortune 500 firms, will now be allowed to use Anthropic's Claude Mythos 5 model without requiring an export licence.
However, access will remain restricted for organisations that are not on the government's approved list.
The move marks a significant shift in Washington's approach after it had directed Anthropic earlier this month to suspend access to some of its most powerful AI models over concerns that they could be used by military or intelligence entities in countries such as China and Russia.
Restrictions Partially Lifted
Following the June 12 export control order, Anthropic had disabled access to its flagship Claude Mythos 5 and Fable 5 models for all users while engaging with the US government to address security concerns.
Advertisement
In a letter to the company, US Commerce Secretary Howard Lutnick said Anthropic had made "significant progress" in mitigating the risks associated with the advanced AI models.
As a result, trusted companies, their foreign-national employees and Anthropic's own foreign-national staff will no longer require export licences to access Claude Mythos 5. Restrictions, however, will continue to apply to companies outside the approved network.
A Commerce Department spokesperson said the government had worked quickly to balance national security with maintaining America's leadership in artificial intelligence.
Fable 5 May Follow
According to a source familiar with the matter, the US government is also considering allowing Anthropic to restore access to its Fable 5 model, although no timeline has been announced.
The easing of restrictions suggests Washington is moving towards a more targeted export-control framework, limiting access only for entities viewed as potential security risks rather than imposing blanket restrictions on advanced AI models.
Advertisement
A Complex Relationship With Washington
Anthropic's relationship with the US government has been under scrutiny in recent months as the company prepares for a potential public listing.
According to Reuters, the AI developer had previously declined to allow its models to be used for domestic surveillance or fully autonomous weapons systems by the US military, contributing to tensions between the company and government agencies.
The latest decision indicates a partial thaw in that relationship, while underscoring the Biden administration's broader effort to safeguard advanced AI technologies from strategic rivals without restricting access for trusted commercial and research partners.