Statement on the US government directive to suspend access to Fable 5 and Mythos 5

US Government Orders Anthropic to Suspend Fable 5 and Mythos 5 Access

The U.S. government has issued an export control directive forcing Anthropic to immediately suspend all access to its new Fable 5 and Mythos 5 AI models for all users, including foreign nationals inside and outside the United States. The abrupt order, which Anthropic received on June 12, cites unspecified national security concerns. This action represents a significant government intervention into the deployment of a commercially available frontier model, raising immediate questions about regulatory standards and the future of AI development.

Anthropic Contests Basis for Suspension

In a public statement, Anthropic explained that the government's concern appears to stem from a reported method of 'jailbreaking' Fable 5. The company claims the technique is a narrow, non-universal jailbreak that identifies minor, previously known vulnerabilities. Anthropic asserts that its 'defense in depth' strategy, which includes extensive pre-launch red-teaming and robust monitoring, was designed with the understanding that perfect jailbreak resistance is not currently achievable by any provider.

• Anthropic stated the demonstrated vulnerability is not unique and can be discovered by other publicly available models, including OpenAI’s GPT-5.5, without requiring a special bypass.
• The company noted that its safeguards were extensively tested for thousands of hours by the US government, UK AISI, and third-party organizations before launch.
• To date, no universal jailbreak has been found for Fable 5, and the company claims the government has not disclosed any evidence of a jailbreak leading to a harmful result.

While complying with the directive, Anthropic strongly disagrees with the action, arguing that applying this standard across the industry would 'essentially halt all new model deployments for all frontier model providers.' The company criticized the lack of a transparent, technically-grounded process and stated it is working to restore access, believing the situation to be a misunderstanding. This incident highlights the growing friction between the pace of AI innovation and the government's evolving approach to managing potential security risks.

This directive marks a shift from regulatory observation to direct, pre-emptive intervention in the AI market. By acting on an undisclosed, narrow vulnerability, the government has introduced a new level of regulatory risk for all frontier model developers. The lack of a clear, technically-grounded framework creates significant uncertainty that could impact investment and deployment timelines across the entire industry.