Anthropic has blocked global access to its Fable 5 and Mythos 5 models following a government order issued Friday evening. The directive cited national security concerns and applied to all foreign nations, foreign nationals within the US, and Anthropic employees themselves.
To comply, Anthropic shut down access entirely rather than implement selective geographic restrictions. This means customers worldwide, including US-based users, lost access to these models simultaneously. The company confirmed the takedown applied universally across all customer segments.
The order represents an escalation in AI export controls. The US has increasingly restricted advanced AI model access abroad, treating frontier models as strategic technologies comparable to semiconductors or military equipment. Fable 5 and Mythos 5 appear to have crossed that threshold.
Details about why these specific models triggered the restriction remain limited. Neither model appears to be Anthropic's flagship Claude product line, suggesting the government targeted particular capabilities rather than the company's entire portfolio. The scope of the ban, however, affects all users indiscriminately.
Anthropic joins other AI companies navigating US export restrictions. OpenAI has faced similar pressures around GPT model availability. These controls reflect broader US strategy to maintain technological advantage while limiting adversaries' access to frontier AI capabilities.
The immediate impact hits paying customers who relied on Fable 5 and Mythos 5 for production workloads. Anthropic typically offers migration paths when discontinuing services, though the government-mandated nature of this shutdown may limit options.
This move underscores the government's willingness to exercise direct control over AI model availability. Rather than working through regulatory frameworks, the administration issued direct orders and Anthropic complied immediately. The precedent suggests future restrictions may come quickly and with minimal warning.
The decision also raises questions about regulatory clarity. AI companies now operate under undefined thresholds for when