The Trump administration has allowed Anthropic to restore access to its powerful Claude Mythos 5 model for selected U.S. organizations, partially reversing an earlier restriction that had blocked access to the company’s top AI systems over national security concerns.
The U.S. government has allowed Anthropic to release Claude Mythos 5 to some trusted U.S. organizations after previously ordering access to be suspended over national security risks.
More Than 100 Organizations Could Get Access
The decision gives Anthropic a path to redeploy Mythos 5, but only to a limited group.
Reuters reported that more than 100 companies and institutions will now have access to Mythos 5, including many Fortune 500 companies, according to a source familiar with the new directive.
Anthropic confirmed the model is being restored for critical infrastructure work. Anthropic said that the government notified the company that Mythos 5, described as its strongest cybersecurity model, can be redeployed to U.S. organizations that operate and defend critical infrastructure.
The company said access will be restored quickly. Anthropic said it is restoring access for those organizations and continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.
A Partial Reversal of an Earlier Shutdown
The move follows an abrupt restriction earlier this month. Anthropic had disabled its most advanced AI models, Mythos 5 and Fable 5, for all users after a June 12 government export control order.
CNBC’s report also framed the update around Anthropic’s Mythos model being released again to selected companies and government agencies.
The U.S. government’s concern centers on potential misuse.
Fears powerful AI systems could be misused by military intelligence users in China, Russia or other countries of concern have pushed the Trump administration toward aggressive oversight of Anthropic and OpenAI’s frontier models.
Export Rules Loosen for Approved Users
The new directive also changes access rules for some non-U.S. citizens. Commerce Secretary Howard Lutnick said an export license will no longer be needed for Mythos 5 access by trusted companies and their employees who are not U.S. citizens, or by Anthropic employees who are not U.S. citizens.
The approved group appears tied to Anthropic’s earlier cybersecurity initiative. Many of the approved companies are part of Anthropic’s Project Glasswing, which includes about 100 well-known tech companies and institutions.
Fable 5 remains a separate issue. Both Fable 5 and Mythos use the same underlying AI model, but Fable 5 is designed for broad public use while Mythos has some safeguards lifted.
Critics Question Government Control Over Access
The selective release is already drawing criticism over transparency. John Coleman, legislative counsel for the Foundation for Individual Rights and Expression, said that, “No one knows how these companies are picked and why everyone else is excluded”.
OpenAI CEO Sam Altman raised a similar concern. Altman wrote on X that extensive safety testing is not a bad idea, but he did not like the idea of the government picking the customers.
Frontier AI Releases Enter a New Phase
The Mythos decision arrives as the U.S. government also pressures OpenAI. OpenAI said it was delaying the full public launch of GPT-5.6 at the U.S. government’s request and limiting access to a small group of vetted partners.
The policy shift shows how frontier AI models are being treated less like ordinary software and more like controlled technologies with national security implications. For Anthropic, the partial Mythos release is a relief. For the wider AI industry, it raises a bigger question: whether future access to the most powerful models will depend not only on product readiness, but also on government approval.
