Anthropic could regain authorization to offer its flagship Fable 5 AI model in the coming days after the Trump administration moves closer to lifting restrictions imposed over national security concerns, Axios reported Saturday, citing sources familiar with the discussions.

The potential turnaround follows a dramatic two-week period. On June 12, the Commerce Department forced Anthropic to take Mythos 5 and Fable 5 offline worldwide despite strong demand from developers, who praised its coding and reasoning capabilities, after a reported jailbreak vulnerability raised cybersecurity concerns.

The thaw in relations comes after the Commerce Department granted limited access to Anthropic’s Mythos 5 cybersecurity model, pointing to improvements in safety risk mitigation and the company’s pledge to work with officials on future AI deployment standards.

Final approval for Fable 5 still depends on the Pentagon and National Security Agency, while both Anthropic and OpenAI continue to push for a standardized regulatory framework governing the release of frontier AI models.