After two weeks of unprecedented blockage, the U.S. government reauthorized on Friday the most powerful artificial intelligence model from Anthropic, for a small circle of only American partners, confirming Washington’s renewed control over this strategic technology.
The blockage had been ordered after Amazon reported a breach in Fable 5’s guardrails, designed to prevent any misuse of the tool.
This conditional restoration comes on the same day that Anthropic’s main rival, OpenAI, launched its new model, GPT-5.6, also with restricted access and client-by-client government validation.
The relaunch of Mythos 5 will for the moment benefit only a small circle of American “cyber defenders and infrastructure operators,” Anthropic announced, saying it is working to restore their access “as quickly as possible”.
Foreign partners, notably state cybersecurity agencies in Europe and Asia, remain off-limits at this stage.
The fate of Fable 5, a consumer version of Mythos limited to cybersecurity and the risks of biological and chemical attacks, also remains unresolved.
The company, whose relations with the Trump administration have been stormy for months, says it is continuing discussions with the government to “broaden access to Mythos 5 and make Fable 5 available to the general public again.”
