Artificial intelligence company Anthropic PBC has received approval from the US government to partially restore access to its advanced Mythos 5 AI model after addressing national security concerns raised by the Trump administration.
According to a letter from Commerce Secretary Howard Lutnick, the company worked with US authorities to mitigate risks associated with its “Covered Models,” leading to what officials described as “significant progress.” The model will now be made available only to select trusted partners.
The decision comes after the US government recently restricted Anthropic from providing foreign nationals access to Mythos 5 and a related system, Fable 5, citing concerns that existing security safeguards could be bypassed. Following the order, Anthropic had temporarily disabled global access to both models while negotiations continued.
A spokesperson for the Commerce Department said the swift resolution reflects efforts to maintain US leadership in artificial intelligence while ensuring national security protections remain intact.
Anthropic confirmed that Mythos 5, its most advanced cybersecurity-focused model, will now be redeployed to a limited group of cyber defence organisations and infrastructure providers. The company said it is working to restore access as quickly as possible for approved users and continues discussions with US authorities regarding broader availability.
While Mythos 5 has received partial clearance, restrictions on the related Fable 5 model remain unchanged, according to the Commerce Department letter.
