Anthropic Pulls Two AI Models to Comply with Trump Administration Security Directive

Anthropic said Friday it will remove access to two of its AI models — Fable 5 and Mythos 5 — to comply with a Trump administration directive restricting foreign nationals from using its latest systems, citing national security concerns.

The move marks a notable intervention by the federal government into the deployment of commercially available AI models, underscoring the growing role of security policy in shaping which users can access frontier AI systems.

Details of the government's request, including how long the restrictions will remain in place, have not been made public, Anthropic said.

The action comes as Anthropic has simultaneously been raising alarms about the broader trajectory of AI development. The company has called for a global pause on the development of powerful AI systems, warning that they are slipping beyond human control and could have catastrophic consequences if not properly regulated.

The dual posture — cooperating with a government security directive while also urging the international community to slow down development — reflects the complicated position Anthropic occupies as both a commercial AI developer and one of the most prominent voices for AI safety.

Anthropic has not specified what prompted the Trump administration's directive, and neither the company nor officials have disclosed which foreign nationals or national groups the restrictions are aimed at.

The removal of Fable 5 and Mythos 5 from general access is significant because both represent Anthropic's latest generation of models, suggesting the administration's concern centers on preventing adversaries from accessing cutting-edge AI capabilities.

The lack of transparency around the directive's scope and duration leaves open questions about how long the restrictions will apply and whether other AI developers could face similar compliance demands.

Anthropic's warning that powerful AI systems risk escaping human control adds a layer of urgency to the policy conversation, even as the company's own newest models become the subject of government-mandated access controls.

For the broader AI industry, the episode signals that Washington is increasingly willing to treat advanced AI models as national security assets subject to export-style restrictions — a dynamic that could reshape how companies develop, deploy, and communicate about their most capable systems going forward.