Anthropic Suspends Claude Fable 5 and Mythos 5 Following US Government Order

Anthropic has abruptly suspended access to its newly released Claude Fable 5 and Claude Mythos 5 AI models following an export control directive from the United States government.

The move comes only days after Anthropic launched Claude Fable 5 and Claude Mythos 5, two next-generation AI models designed for different audiences. While Fable 5 targets general users with stricter safety protections, Mythos 5 focuses on advanced coding, cybersecurity, and research capabilities. Anthropic had also recently introduced Project Glasswing with Claude Mythos Preview, keeping its most advanced capabilities out of public hands.

According to Anthropic, the government order requires the company to block access to both models for all foreign nationals. The restriction applies regardless of whether those individuals are located inside or outside the United States. It also affects foreign national employees working at Anthropic.

To ensure compliance, Anthropic disabled access to both models for all customers worldwide while it evaluates the directive and works with regulators.

Government Order Cites National Security Concerns

Anthropic said the export control directive did not provide detailed information about the specific national security risk that prompted the action.

However, the company believes the concern may be linked to a potential jailbreak technique affecting Claude Fable 5. A jailbreak is a method that attempts to bypass an AI model’s built-in safety restrictions.

The company emphasized that no universal jailbreak has been discovered for Fable 5 and argued that the issue appears limited in scope.

Anthropic Defends Safety Testing Efforts

Anthropic said it invested significant resources into securing Claude Fable 5 before release.

According to the company, the model underwent thousands of hours of safety evaluations involving internal teams, third-party organizations, government partners, and the UK’s AI Safety Institute (AISI).

The company said it specifically designed Fable 5 with safeguards intended to prevent misuse for cybersecurity and offensive security activities.

Anthropic added that extensive red-team testing failed to uncover any universal method for bypassing the model’s protections.

Older Models Remain Available

While access to Fable 5 and Mythos 5 has been suspended, Anthropic confirmed that older Claude models remain available and are not affected by the export control order.

The company warned that applying similar standards across the broader AI industry could significantly slow the deployment of frontier AI systems.

Anthropic argued that if companies are forced to withdraw models over theoretical or narrowly scoped vulnerabilities, future AI launches could face major delays or become economically impractical.

Microsoft Previously Restricted Fable 5 Access

The suspension follows another recent setback for Claude Fable 5.

Earlier, Microsoft reportedly restricted employee access to the model because of concerns surrounding Anthropic’s updated data retention policies.

Microsoft’s concerns were reportedly unrelated to the newly announced export control directive.

Anthropic said it is actively working with authorities to restore access to Claude Fable 5 and Claude Mythos 5 as quickly as possible.

According to the latest information, Anthropic’s Fable 5 and Mythos AI models were banned following Amazon’s warning, and the company is working on addressing the security concerns.

For now, customers attempting to access either model will be unable to do so until the restrictions are lifted.