Anthropic AI company has restricted access to its new AI models, Fable 5 and Mythos 5, after concerns from US authorities.
The company announced the decision on Friday and said it needed to follow export rules and government requirements.
Anthropic, foreign nationals can no longer use Claude Fable 5. The company described the model as extremely powerful and said it had to take immediate action to remain compliant with regulations.
Anthropic also disabled Fable 5 and Mythos 5 for all customers for a short period. The company said the move would help ensure compliance with the government order. However, it confirmed that its other AI models remain available and are not affected.
The announcement came only days after Anthropic introduced the two new models. The company promoted them as some of its most advanced AI systems and said they performed strongly across several industry tests.
Claude Fable 5 is a protected version of the Claude Mythos engine. The models represent an important step for Anthropic as it competes with major AI companies such as Open-AI and Google.
Anthropic said this is the first time it has launched such advanced models while using stronger safety protections. The company developed these protections to reduce risks and limit harmful uses of AI technology.
The company first introduced its restricted development strategy through a cyber-security program known as “Project Glass wing”. Later, Anthropic said the government did not provide detailed information about the national security concerns behind the order.
Even so, the company stated that governments should have the power to stop unsafe AI deployments. At the same time, Anthropic said any such process should remain transparent, fair, clear, and based on technical evidence.
The latest development marks another disagreement between Anthropic and the US government. Earlier this year, the company faced challenges in its relationship with the Department of Defense.
After talks between the two sides ended without an agreement, the Department of Defense classified Anthropic as a supply chain risk. This designation created additional restrictions for defense contractors.
As a result, contractors working with the US military must confirm that they do not use Anthropic Claude models in certain projects. The label has often been used in cases involving serious security concerns.
