US Bars Foreign Nationals From Anthropic's Top AI Models

Anthropic announced on Friday that it had disabled customer access to its Fable 5 and Mythos 5 artificial intelligence models following a directive from US government agencies. The order cited national security concerns and required the company to block usage by all foreign nationals, including those working inside the United States and even within Anthropic itself. Other company products, such as its Claude chatbot, remain unaffected.

The directive arrived at 5:21 p.m. on June 12. Anthropic stated that the letter provided no detailed explanation of the specific risks involved. To meet the requirement, the firm suspended access across its entire customer base rather than attempting selective restrictions. This step halted work for paying users as well as internal staff who hold foreign citizenship.

Fable 5 was released publicly on June 9 and incorporates many capabilities previously limited to Mythos, a model restricted to select Project Glasswing partners. Anthropic described Fable 5 as exceeding prior models in performance, noting that it completed tasks such as beating Pokémon FireRed in company tests where earlier versions had fallen short. Mythos 5 focuses on cybersecurity functions, including the detection of long-undiscovered software vulnerabilities. Those same functions have been employed by US agencies and approved companies to address security gaps.

The government action appears linked to reports of a jailbreak method that could circumvent safeguards on Fable 5. Anthropic had previously outlined steps taken during development to limit misuse, yet the directive overrode those internal controls. The broad scope of the order, which treats foreign nationals inside the country the same as those abroad, produced an immediate operational shutdown for many users regardless of location or employment status.

Export control rules have long applied to sensitive technologies, yet their extension to newly released AI systems raises practical questions about enforcement and scope. Companies must now weigh rapid compliance against the risk of disrupting legitimate research and commercial activity. In this case, the absence of granular guidance left Anthropic with little choice but to implement a blanket cutoff.

Market participants have responded by exploring alternative providers or internal workarounds while awaiting further clarification. The episode illustrates how regulatory decisions can quickly alter the availability of tools that private firms develop through iterative testing and customer feedback. Historical patterns in technology policy show that such interventions often expand beyond initial targets and carry costs that fall on domestic users as well as foreign ones.

Anthropic continues to operate its remaining models and services. The company has not indicated when or whether access to Fable 5 and Mythos 5 might resume under revised conditions. Government agencies have offered no additional public statements on the precise nature of the national security concerns or the timeline for any review process.