The Trump administration appears to be moving towards restoring full access to Anthropic’s Fable 5 model, which has been offline for 15 days following an abrupt government intervention that sent shockwaves through the developer community and forced enterprises to scramble for alternatives.
The partial thaw came on Friday, when Commerce Secretary Howard Lutnick wrote to Anthropic confirming that Mythos 5, the company’s most powerful cybersecurity-oriented model, could be made available to approximately 100 pre-approved US government agencies and companies. Fable 5, the public-facing version of the same model class, remains restricted, but a source close to the situation told Axios that its return could come as early as this week.
The standoff began on June 12, when Lutnick sent a letter to Anthropic CEO Dario Amodei placing both models under export controls after another company claimed to have jailbroken Mythos 5, raising national security concerns within the administration. The government had attempted to persuade Anthropic to pause the models’ release voluntarily, but was unsuccessful. Export controls followed, and Anthropic cut off access to both models for all customers, including its own non-American employees, to comply.
The move was unprecedented in scope. Developers found automated workflows frozen mid-task, and enterprises pivoted rapidly to rival models, including cheaper Chinese alternatives. Stripe had been using Fable 5 to overhaul a 50-million-line codebase in a single day, a job that would have taken its engineers more than two months by hand.
The dispute sits within a broader and more serious conflict between Anthropic and the US government. After contract negotiations with the Pentagon broke down, the Department of Defense declared Anthropic a supply chain risk to national security, a designation historically reserved for foreign adversaries, effectively barring defence contractors from using Claude models in military work. Anthropic launched legal action against the administration to reverse the blacklisting, and litigation remains ongoing. Tom Brown, Anthropic’s co-founder, has since taken over negotiations with the administration from Amodei.
Friday’s directive marked a notable change in tone. “Anthropic has worked positively with the government,” one administration source told Axios, a striking shift from Defence Secretary Pete Hegseth’s earlier public designation of the company as a national security risk.
Whether Fable 5 returns with the same terms users had before its withdrawal remains unclear. Axios reported that it is not yet known whether subscribers will regain unrestricted access or whether the model returns behind additional fees or identity verification requirements. The Pentagon and the National Security Agency have yet to formally clear Fable 5, meaning the final outcome remains uncertain.
Both Anthropic and OpenAI are now pushing the administration to establish a formal, codified process for reviewing new models, as envisioned by a June 2 executive order setting up a voluntary government vetting framework, rather than continuing with the current case-by-case approach.
RELATED STORIES
Anthropic Mythos test finds flaws in classified US government systems within hours, reports say
Anthropic takes Project Glasswing to critical infrastructure operators in 15 countries
