Anthropic cleared to release Mythos AI to 100 US firms
Anthropic said the US government allowed it to make its powerful Claude Mythos 5 AI model available to some “trusted” US organizations.
A source familiar with the new directive, whose name was not disclosed due to the sensitivity of the issue, said that more than 100 companies and institutions, including Fortune 500 companies, can now access Mythos 5.
Concerns that powerful AI systems could be misused by military intelligence users in China, Russia or other worrisome countries have prompted President Donald Trump’s administration to take an aggressive approach to surveillance of releases of frontier models from Anthropic and rival OpenAI.
Anthropic abruptly disabled its most advanced AI models (Mythos 5 and Fable 5) for all users following the government’s June 12 export control order.
“Today, the government notified us that Mythos 5, our most powerful cybersecurity model, may be redeployed to a number of U.S. organizations that operate and defend critical infrastructure,” Anthropic said in a statement. he said.
“We are quickly restoring access to these organizations and continue to work with the government to expand access to Mythos 5 and bring Fable 5 back to the general public,” he added.
The government’s review of which companies can access Mythos has drawn much criticism.
“Nobody knows how these companies were singled out and why everyone else was excluded,” said John Coleman, legislative counsel for the Foundation for Individual Rights and Expression, a Philadelphia-based nonpartisan free speech organization.
“This puts too much power in the hands of the government. There is too little transparency and it raises questions about the rule of law.”
In a post on X, OpenAI boss Sam Altman reiterated concerns about the government’s choice of who has access to the best models.
Comprehensive security testing “is not a bad idea. I just don’t like the idea of the government picking customers,” he wrote.
Experts say that if Mythos models fall into the wrong hands, they can significantly accelerate sophisticated cyberattacks, especially in industries like banking that rely on complex, interconnected and often decades-old technology systems.
A letter to Anthropic from US Commerce Secretary Howard Lutnick said the company had made “significant progress” in its work with the government to address “risks associated with Covered Models.”
It was not immediately clear what precautions were taken.
Lutnick said in the letter that export licenses for Mythos 5 to trusted companies and their non-U.S. citizen employees or to Anthropic’s non-US citizen employees will no longer be required, but licensing restrictions will remain in place for companies not on the approved list.
Most of the approved companies are part of Anthropic’s Project Glasswing, which includes around 100 well-known technology companies and institutions, the source said.
The source said the government is also moving towards allowing Anthropic to publish Fable soon, but the timeline is unclear.
