Anthropic's strongest AI is making a return, but exclusively for a limited group.

      The U.S. government has given the green light for the limited reinstatement of Mythos 5 as Fable 5 approaches a broader rollout.

      Anthropic's AI restrictions may be easing. Following an unexpected shutdown earlier this month due to U.S. government security issues, the company's most advanced AI models are gradually returning. A recent Axios report indicates that Anthropic has already reinstated access to Mythos 5 for a small group of trusted users, while Fable 5 could be back as early as next week if talks with federal agencies remain productive.

      Mythos is reintroduced first, as Fable waits for approval.

      Axios reports that the U.S. Commerce Department has authorized Anthropic to restore Mythos 5, its premier AI model developed for cybersecurity, to a select group of trusted clients. Unlike Fable 5, Mythos was never made widely available to the public and incorporates additional safeguards to mitigate the risks of misuse in areas like cyberattacks and biological threats.

      The main objective, however, is Fable 5. Axios notes that the Trump administration is on the brink of lifting restrictions on the model after nearly two weeks of discussions with Anthropic. The model was pulled shortly after its launch, despite quickly establishing itself as one of the most capable AI systems in the industry, especially in coding and deep reasoning. Developers who had already integrated Fable 5 into their processes were left in a bind when access was abruptly revoked.

      Commerce Secretary Howard Lutnick reportedly informed Anthropic that the company had made considerable headway in addressing federal concerns about both Mythos 5 and Fable 5, with Anthropic agreeing to continue collaborating with U.S. agencies on future release protocols. Nevertheless, Axios emphasizes that final approval for Fable 5 still hinges on additional agencies, such as the Pentagon and the National Security Agency.

      This situation reflects an emerging norm for advanced AI systems.

      Interestingly, the reinstatement of Mythos 5 reflects broader trends in AI regulation rather than just Anthropic's individual case. The company initially marketed Fable 5 as its primary public model, while Mythos 5 serves as a more stringently controlled alternative, equipped with safeguards for high-risk scenarios. Allowing Mythos to return first implies that regulators are becoming more accepting of approving AI systems with stricter controls before permitting their more open counterparts to operate freely.

      The larger trend shows that this is no longer merely an issue for Anthropic. OpenAI has adopted a remarkably similar approach with its recent GPT-5.6 preview, restricting access to trusted partners while undergoing the same governmental review processes. A clear trend is emerging: the launch of the world’s most advanced AI models has evolved beyond just a technical achievement; it has become a regulatory one too.