Anthropic Fable 5 could return: US moves to lift ban

Anthropic's Fable 5 could be available again within days, Axios reports, after the Trump administration moved to lift restrictions that forced the model offline. Commerce Secretary Howard Lutnick told Anthropic in a letter that the company had "worked with the US government to address risks." The Pentagon and NSA still need to sign off.

What's happening now?

Fable 5 may return to users imminently because the Trump administration is close to lifting restrictions imposed on the model, though the Pentagon and NSA must still approve the decision. The possible return follows a government-ordered shutdown and comes as Anthropic and OpenAI push for a legally defined review process for new AI models rather than case-by-case decisions.

Anthropic's latest AI model was taken offline on June 12 after US government authorities ordered its shutdown over safety concerns. Axios, cited in reporting on June 27, says the administration is preparing to lift those limits, which would allow Fable 5 to be made available again in the coming days if the remaining agencies sign off.

How did Fable 5 get pulled and what else changed?

Fable 5 was shut down on June 12 on orders from the US government over safety concerns, and Anthropic also took Mythos 5, a variant without added safety restrictions, offline at the same time; Mythos 5 is already back for select partners. The shutdown was a direct government action; the return hinges on interagency approval.

Anthropic faced a simultaneous removal of both Fable 5 and Mythos 5. The Decoder notes Mythos 5, described as the variant without Anthropic's added safety restrictions, has already returned for a limited set of partners. That contrast — one model back for partners, the other still restricted pending sign-off — raises immediate questions about what conditions or safeguards will accompany any reopening of Fable 5.

Commerce Secretary Howard Lutnick communicated with Anthropic in a letter, saying the company had "worked with the US government to address risks," a phrase the administration appears to treat as evidence that changes made by Anthropic meet the government's concerns. The Pentagon and NSA remain the outstanding approvers whose clearance the administration expects before lifting the broader restriction.

Why it matters

A clearance for Fable 5 would show the US government moving from unilateral takedowns toward an approval path for advanced models, and it sharpens the debate over how model releases should be governed. Both Anthropic and OpenAI are pushing for a legally defined review process for new models instead of ad hoc decisions, a shift that would redraw how major models reach users and could change commercial timelines and partner arrangements.

The distinction between Mythos 5 returning for select partners and Fable 5 remaining restricted highlights a second practical point: regulators may approve narrower or conditional access instead of full, worldwide releases. OpenAI also expects full clearance for its own model only in the coming weeks, suggesting multiple companies are facing staggered, agency-dependent reopenings rather than instant, universal rollouts.

What to watch

Watch for explicit sign-off from the Pentagon and the NSA; their approvals are the immediate gating items named in the reporting. Also watch whether any clearance includes ongoing restrictions similar to those applied to GPT-5.6 Sol, or whether Fable 5 will be permitted to resume worldwide access, and whether Congress or industry actors push a legally defined review process into law.

June 12 (shutdown) and the June 27 reporting together frame the immediate timeline for the story: a government-ordered takedown followed by a potential reversal within days if outstanding agencies approve the changes Anthropic made.