Two weeks after the initial ban led Anthropic to withdraw its advanced cybersecurity models, Mythos 5 and Fable 5, the Trump administration is beginning to ease restrictions. The administration has granted Anthropic permission to make Mythos 5 available to over 100 specific U.S. government agencies and companies. Notably, this approval extends to include non-American employees within these organizations, a reversal from the original ban that prohibited non-U.S. residents from accessing the model.
Commerce Secretary Howard Lutnick communicated this decision in a letter to Anthropic’s chief compute officer, Tom Brown, stating that sufficient safeguards are now in place to allow trusted partners access to Mythos 5.
However, it is important to note that the directive did not mention the availability of the Fable 5 model, which had been released just days before the ban and was promoted as having enhanced security measures. Both models were withdrawn after reports indicated that security researchers were able to bypass their protective features.
In a public statement on social media platform X, Anthropic acknowledged the advancements made in discussions with the U.S. government. The company highlighted that since June 12, it has been actively working to restore access to its models. The recent notification allows Mythos 5, described as Anthropic’s strongest cybersecurity model, to be redeployed to U.S. organizations responsible for maintaining and protecting critical infrastructure. Anthropic is committed to quickly restoring access to these organizations and is continuing efforts to expand the availability of Mythos 5, alongside plans to reintroduce Fable 5 for broader usage.
