The Trump administration removed export controls on Anthropic’s Fable 5 (also referred to as Claude Fable 5) on Tuesday evening, restoring public access to the company’s Mythos‑class model. Access to the model had been suspended 18 days earlier for security reasons.
What happened and when
- The federal government lifted the export controls on Tuesday evening, ending restrictions that had limited availability of Fable 5.
- The model’s access had been pulled 18 days prior due to security concerns.
- Last week the administration allowed Anthropic to restore access to Mythos 5 for a limited set of government‑approved organizations.
Why it matters
Restoring access to Fable 5 matters for AI developers and users because the model belongs to Anthropic’s Mythos tier, representing high capability; its availability affects research, development and competitive dynamics in the field. The episode also underscores ongoing interactions between government security concerns and companies’ deployment plans.
Government and industry context
- The Commerce Department had expressed concern about preventing access by foreign nationals; the specific technical or policy changes Anthropic made to address those concerns were not disclosed publicly.
- Commerce Secretary Howard Lutnick posted on X that his office had “worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America's leadership in AI.”
- Similar, limited rollouts have occurred elsewhere in the industry: OpenAI made GPT‑5.6 available only to a small set of approved customers last week following a request from the U.S. government.
Risks and geopolitical implications
U.S. officials, allies and major AI companies have grown more worried that frontier models could be misused—to automate sophisticated cyberattacks or speed development of biological threats. At the same time, China is advancing toward producing a competing Mythos‑class system while continuing to release open‑weight models that rival systems like Anthropic’s Claude Opus 4.8.
What’s next
Under a recent executive order, the Trump administration has an August deadline to create standardized benchmarks for evaluating the security risks of new AI models. The handling of Fable 5 illustrates that U.S. policy on pre‑release review and control of frontier models remains in flux.
