Two weeks after the ban that led to Anthropic withdrawing its robust cybersecurity-focused models, Mythos 5 and Fable 5, from the marketplace, the Trump administration is easing its position.

The administration is now permitting Anthropic to offer Mythos 5 to over 100 designated U.S. government agencies and companies, which includes access for the non-American staff of those entities, as reported by both Semafor and Reuters. This roster also encompasses Anthropic’s own non-American personnel, who were part of the initial prohibition that restricted non-Americans from using the models.

“I have concluded that suitable precautions are in place to allow select trusted partners to utilize the Claude Mythos 5 Model,” Commerce Secretary Howard Lutnick communicated to Anthropic’s chief compute officer Tom Brown on Friday, according to the letter reviewed by Semafor.

It appears that the administration did not mention the release of Fable 5 in this communication. This version of Mythos 5 was broadly launched a few days prior to the ban as it was claimed to have enhanced protections. Both models were retracted after those safeguards were reportedly easily circumvented by security researchers. Anthropic did not quickly respond to our inquiry for comment.

On Friday, Anthropic publicly recognized the advancement in a statement on X, stating: “Since June 12, we’ve been collaborating closely with the U.S. government to regain access to Claude Mythos 5 and Fable 5. Today, the government informed us that Mythos 5, our most powerful cybersecurity model, can be redeployed to a group of U.S. organizations that maintain and protect critical infrastructure. We’re expediting access restoration for these organizations and are continuing to collaborate with the government to broaden access to Mythos 5 and make Fable 5 available for public use once again.”