Anthropic announced an emergency shutdown of access to AI models Fable 5 and Mythos 5 for all users after receiving a directive from the US government export controls. The directive, issued with reference to national security authorities, prohibits access to these models by any foreign nationals, both inside and outside the United States, including foreign employees of Anthropic. This has resulted in the company being forced to completely disable the models to ensure compliance, although access to other Anthropic models remains unaffected.
The reason for the regulation, according to Anthropic, was the discovery of a method to bypass the protective mechanisms of the Fable 5 model, known as jailbreaking. Anthropic stated that the technique is a narrow, non-universal bypass that only identifies a small number of previously known vulnerabilities that can be found in other publicly available models, such as GPT-5.5, without any workaround.
While the company has complied with the legally binding directive, Anthropic does not agree that the discovery of a narrow potential bypass justifies revoking access to a commercial model deployed to hundreds of millions of users. They consider the incident a misunderstanding and are working to restore access to Fable 5 and Mythos 5.