Anthropic suspended its new Claude Fable 5 and Mythos 5 models just days after their public launch.
The company said it had received an order to restrict foreign users’ access to Claude Fable 5. As a result, Anthropic disabled both models for all customers.
According to the company, U.S. authorities believe a way had been found to bypass the model’s built-in safety mechanisms. The details of the claims have not been disclosed.
Anthropic says the issue concerns a small number of already known vulnerabilities. The company also said similar problems can be identified by other publicly available AI models without using special methods to bypass protections.
Anthropic had previously described Claude Fable 5 as its most powerful model. According to the company, its capabilities in finding vulnerabilities and working with computer systems surpassed its earlier developments.
Against this backdrop, a legal dispute between Anthropic and the administration of Donald Trump continues. Pentagon chief Pete Hegseth earlier called the company a “risk to supply chains,” after which Anthropic filed a lawsuit against the department.
Anthropic and OpenAI AI Models Have Become Much Stronger at Finding Vulnerabilities
White House Sought to Remove Anthropic From U.S. Government Systems but Is Now Forced to Work With the Company Again
Anthropic’s AI Found Thousands of Zero-Day Vulnerabilities in the Largest Operating Systems and Browsers
Anthropic Sued the Trump Administration After the Pentagon Labeled the Company a Supply Chain Threat