U.S. Government Imposes Sudden Export Controls, Blocking Anthropic Fable 5 and Mythos 5 Models

The U.S. government recently issued an emergency directive to AI company Anthropic, ordering it to immediately suspend access to its two new models, Fable 5 and Mythos 5, for non-U.S. users. This move is seen as the latest export control measure targeting AI models, citing national security and potential cyber capability risks, and has drawn significant global attention from the tech industry.

According to sources familiar with the matter, the control action occurred less than 48 hours after the models' official release. Anthropic subsequently issued a statement on its website confirming that it had restricted overseas access as required by the government. The statement did not specify any particular country, but industry observers widely believe the move is related to the recent escalation of tech tensions between the U.S. and China.

Fable 5 and Mythos 5 are a new generation of reasoning models launched by Anthropic in early 2025, claiming breakthroughs in multimodal understanding and long-context processing. Shortly after their release, they attracted a large number of overseas developers for testing, but access was cut off within just two days, sparking strong backlash from users.

In an internal memo, the Bureau of Industry and Security (BIS) of the U.S. Department of Commerce noted that the two models possess "potential dual-use capabilities" that could be used for cyberattacks or military purposes if they fell into the hands of certain countries. The memo emphasized that the models' weight files and inference interfaces fall under the category of controlled technology.

After the news broke, related topics on platform X quickly trended, with discussion volumes exceeding one million. Many users compared this incident to previous access restrictions on models like ChatGPT and Grok, arguing that AI is transforming from a "general-purpose tool" into a "strategic resource."

Analysts point out that the ban reflects the U.S. government's growing concern over the proliferation of cutting-edge AI models, extending from "chips" to "the models themselves." Traditional export controls mainly targeted semiconductor hardware, but now algorithms and training data are also falling under scrutiny.

Industry experts warn that unilateral export restrictions may accelerate the fragmentation of the global AI ecosystem. Some European and Asian countries have begun discussing the establishment of domestic model training infrastructure to reduce reliance on U.S. companies.

For Anthropic, the incident is undoubtedly a major blow. The company had previously secured multiple rounds of massive funding, with its valuation once exceeding $18 billion. The forced removal of the models not only impacts business revenue but may also damage its reputation within the global developer community.

Legal experts say the legality of such administrative directives remains controversial. Some lawyers argue that the government needs to provide more substantial evidence that the models pose an "imminent threat," or it could face judicial challenges.

Currently, Anthropic has established a dedicated team to communicate with the government, hoping to restore some overseas access within a compliant framework. However, in the short term, full access to Fable 5 and Mythos 5 for non-U.S. users is impossible.

This incident once again highlights the complex role of AI technology within the national security framework. In the future, global AI companies may need to plan for multiple versions of their models in advance to cope with increasingly stringent cross-border data and technology flow controls.