Anthropic Fable 5 and Mythos 5 were suspended after the U.S. government issued an order using national security authorities that required the company to block foreign access, Anthropic said June 12.
In a written statement, Anthropic said the U.S. government instructed the company to temporarily suspend access to the Fable 5 and Mythos 5 models for foreign users, including foreign employees located in the United States. The company said other Anthropic models remain available while it works to comply with the directive.
Why Anthropic Fable 5 was suspended
Anthropic said the government did not provide detailed national security concerns when it gave the order. The company said it understands the action may be linked to a specific technique that bypasses Fable 5 safety filters to reveal a small set of known, low-severity software vulnerabilities.
Anthropic added that similar capabilities can be reproduced with other public models, and the company said it has not seen a universal jailbreak that would broadly defeat protections across systems. The company told customers it paused Fable 5 and Mythos 5 to ensure compliance while it seeks clarification from authorities.
Industry reaction and reporting
U.S. outlets Axios, The Verge, and Tom’s Hardware reported on the suspension and the government order. Those outlets cited Anthropic’s statement and people briefed on the matter, and they described the move as an unusual use of export control and national security authorities for cloud-hosted AI models.
Security researchers and corporate customers told reporters they were surprised by the scope of the order, because it restricts model access based on user nationality rather than the models’ code or feature set. Companies that rely on API access or enterprise versions will need to reassess contingency plans for models such as Anthropic Fable 5 and alternatives.
What it means for customers
Teams that integrate models via API or enterprise software can no longer evaluate vendors only on speed and price, industry lawyers and consultants said. They must also determine fallback models, data retention policies, and whether their operations can withstand sudden geopolitical intervention.
Anthropic emphasized that the suspension is a compliance measure, not a product safety recall. Still, the abrupt pause of Anthropic Fable 5 underscores a broader shift: the availability of leading-edge AI models is now part of a product risk calculus for businesses and government users alike.
Companies planning deployments should document substitution plans and review contractual protections for service interruptions, legal experts said. The episode also raises questions about how governments will treat advanced AI models in future export control or national security reviews.
Anthropic did not immediately provide additional public details beyond its June 12 statement, and the company said it is seeking further guidance from U.S. authorities. The situation remains fluid, and affected customers should monitor official communications from Anthropic for updates.
