Anthropic Disables Fable 5 And Mythos 5 Worldwide Following US Government Directive

Artificial Intelligence company Anthropic has suspended access to its Fable 5 and Mythos 5 models after receiving an export control directive from the United States government. The order requires the company to block access for all foreign nationals, both inside and outside the United States, including foreign employees working at Anthropic.

According to the company, the directive was received on June 12 and took immediate effect. As a result, Anthropic said it had no option but to disable the two models for customers worldwide in order to comply with the legal requirement. Access to the company's other AI models remains unaffected.

Anthropic stated that the government did not provide detailed information about the national security concerns behind the decision. However, the company said it believes the action is related to concerns over a method that could bypass, or "jailbreak," certain safety protections built into Fable 5.

The company acknowledged reviewing demonstrations of the reported technique and said it was able to identify a small number of previously known vulnerabilities. Anthropic described those issues as limited in scope and said similar results could be obtained using other publicly available AI systems.

Before launching Fable 5, the company said it conducted extensive safety testing with government agencies, independent organizations, and internal security teams. According to Anthropic, those evaluations involved thousands of hours of testing designed to assess the effectiveness of the model's safeguards against misuse, including cybersecurity-related risks.

Anthropic maintained that it has not been presented with evidence of a broad method capable of bypassing the model's safety protections across a wide range of functions. The company also said it had not received reports linking the identified vulnerabilities to any harmful real-world outcome.

The company emphasized that it had implemented multiple layers of security protections, including monitoring systems and data-retention measures intended to detect and respond to attempts to circumvent safeguards. Anthropic stated that it continues to support this approach and believes it remains an effective method for managing potential risks.

While complying with the government directive, Anthropic expressed disagreement with the decision to suspend the models. The company argued that the concerns raised involve a limited vulnerability rather than a widespread security failure and said it believes regulatory actions affecting advanced AI systems should be based on transparent processes and clearly defined technical evidence.

Anthropic said it is working with relevant authorities and hopes access to Fable 5 and Mythos 5 can be restored in the future. The company also apologized to customers affected by the sudden suspension.