US Approves Anthropic’s Claude Mythos 5 for Trusted US Organizations

The US government has granted Anthropic permission to release its advanced AI model, Claude Mythos 5, to certain “trusted” organizations. This decision reverses a previous order from two weeks ago which suspended access due to national security concerns.

Over 100 companies and institutions will now have the ability to use Mythos 5, including many Fortune 500 firms. The move comes after President Donald Trump’s administration took an aggressive stance on oversight of Anthropic’s and rival OpenAI’s frontier models following fears that powerful AI systems could be misused by military intelligence users in countries like China or Russia.

OpenAI had previously delayed the full public launch of its GPT-5.6 model at the government’s request, limiting access to a small group of vetted partners. Anthropic had abruptly disabled its most advanced models – Mythos 5 and Fable 5 – for all users after the government’s June 12 export control order.

Anthropic stated that it is restoring access for these organizations quickly and continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again. However, the vetting process by which companies are selected has drawn criticism from experts and free speech advocates alike.

OpenAI boss Sam Altman echoed concerns about the government’s choice of who gets access to top models in a post on X. He wrote that extensive safety testing is not a bad idea but he doesn’t like the idea of the government picking customers.

Experts have warned that Mythos models, if used improperly, could significantly accelerate sophisticated cyberattacks, particularly in sectors such as banking which rely on complex, interconnected and often decades-old technology systems.

The US Commerce Secretary Howard Lutnick said there had been significant progress in work done by Anthropic with the government to address risks associated with the Covered Models. However, it is unclear what safeguards have been adopted.

Anthropic stated that it understood the government believed there was a method of bypassing or “jailbreaking” a safeguard that would prevent Fable 5 from being used in identifying software vulnerabilities. The source said many of the approved companies are part of Anthropic’s Project Glasswing, which includes about 100 well-known tech companies and institutions.

The government is also moving towards allowing Anthropic to release Fable soon, although a timeline has not been provided. Both Fable 5 and Mythos use the same underlying AI model but Fable 5 is designed for public use while some safeguards are lifted for Mythos.

This latest move by the US government follows an executive order this month establishing a voluntary framework for AI developers to offer “covered frontier models” to the US government for up to 30 days before releasing them to trusted partners. However, analysts say that without a system in place for companies to widely release new models, it increases the likelihood that China will catch up.