Anthropic suspends access to its latest AI models, Claude Fable 5 and Mythos 5, after concerns were raised by US national security authorities.
The company announced that it had been ordered to restrict foreign nationals from using Claude Fable 5. As a result, Anthropic decided to disable both Fable 5 and Mythos 5 for all users to meet compliance requirements.
According to Anthropic, the government did not provide details about the specific security risks. However, the company believes officials were concerned about a possible method of “jailbreaking” the AI model.
Jailbreaking is a technique used to bypass software restrictions. It can allow users to access hidden features or identify system vulnerabilities.
Anthropic said it reviewed a demonstration of the reported jailbreak method. The company found that it only exposed a small number of previously known and minor vulnerabilities.
It also noted that other publicly available AI models could identify the same issues without special bypass techniques.Before launching Claude Fable 5, Anthropic highlighted several security safeguards designed to reduce cyber risks.
The model was initially released to a limited group of organizations in April for testing and evaluation. At the time, Anthropic described Fable 5 as its most capable AI system yet.
The company had even suggested the model was almost “too powerful to release,” a claim that sparked debate among critics who viewed it as mere marketing hype.
Meanwhile, experts say the suspension could slow down important research and testing efforts. Professor Gina Neff of Queen Mary University London warned that restricting access may limit collaboration between companies and governments working on AI safety.
She also pointed to testing by the UK AI Security Institute, which reportedly found the model could successfully exploit defenses and systems 73% of the time.
The suspension comes as Anthropic faces growing pressure from the Trump administration.
The company is currently involved in a legal dispute with The Pentagon after being labeled a “supply chain risk.” A federal judge has temporarily blocked enforcement of that designation while the lawsuit continues.
