Anthropic announced on July 1 that it has resumed global service of its latest AI model, Claude Fable 5. The security-focused Claude Mythos 5 has also been restored for US organizations with government approval.

Timeline

Fable 5 and Mythos 5 launched on June 9, but on June 12, Amazon researchers discovered a jailbreak method that allowed Fable 5 to identify software vulnerabilities. The US Commerce Department immediately imposed export controls, suspending both models worldwide.

Anthropic worked with the US government to implement enhanced safety measures, leading to the lifting of controls on June 30. Services are being restored in phases starting today.

Availability

Fable 5 is now accessible through Claude.ai, the Anthropic Platform, and Claude Code. Mythos 5 is restricted to US organizations with government approval, including defense and critical infrastructure entities.

Industry Response

While AI security experts have largely praised the response, some caution that similar vulnerabilities may emerge in the future. Anthropic CEO Dario Amodei stated the company remains “committed to balancing safety and innovation.”