The Trump administration has ordered restrictions on foreign citizens' access to Anthropic's most powerful AI models, Fable 5 and Mythos 5. To comply with export controls and national security requirements, Anthropic has been forced to disable access to these models for all clients worldwide who do not hold US citizenship.
What Happened
Anthropic has disabled access to its most advanced models, Fable 5 and Mythos 5, for foreign citizens. The ban applies to both users outside the US and foreign citizens located within the country, including Anthropic's own employees. The decision was made as part of the implementation of a US government directive to ensure national security.
Context
The Fable 5 and Mythos 5 models (including specialized cybersecurity features in Mythos 5) are classified as strategic US national security assets. This move reflects the growing role of export controls in the high-tech sector, where advanced AI systems are viewed as tools of geopolitical influence.
Why It Matters for the Industry
For the industry, this sets a precedent where access to SOTA (State-of-the-Art) models may be restricted not only by geography but also by the citizenship of users and personnel. This leads to the fragmentation of the global AI market into technological blocs, stimulates demand for open-weights solutions, and forces companies to account for political risks when designing global ML infrastructures.
Why It Matters for Users
Foreign specialists and international companies may face an immediate loss of access to key development tools, leading to the disruption of existing workflows. Users outside the US will need to urgently switch to alternative solutions, such as models from OpenAI or open-source models.
What Is Not Yet Known / Limitations
The exact criteria for verifying citizenship for API access and the long-term consequences for distributed international teams remain to be clarified.
Sources
Author
Look at AI, Editorial Team