Anthropic announced that Claude Fable 5 will return to global availability on July 1, following US export control restrictions imposed on June 12. The company suspended access to both Fable 5 and Mythos 5 models after the government flagged concerns about potential cybersecurity vulnerabilities. Now, Fable 5 will be available across the Claude platform with safeguards in place for Pro, Max, Team, and select Enterprise plans.

The restrictions stemmed from an Amazon research report identifying a method to bypass Fable 5's safeguards by prompting the model to identify software vulnerabilities and produce exploit code. Anthropic's testing revealed that several other models—including Claude Opus 4.8, GPT-5.5, and Kimi K2.7—could perform the same tasks, indicating the issue wasn't unique to Fable 5. The reported technique involved routine defensive cybersecurity work rather than exposing Mythos-level capabilities.

Anthropic responded by training an improved safety classifier that blocks the specific bypass technique in over 99% of cases. The new system flags potentially harmful cybersecurity requests and redirects them to Opus 4.8 instead. However, this enhanced protection comes with tradeoffs—users may experience more false positives during legitimate coding and debugging tasks as the classifier errs on the side of caution.

Moving forward, Anthropic is collaborating with Amazon, Microsoft, and Google to develop a shared industry framework for assessing and addressing AI model jailbreaks. This initiative aims to create consistent standards for evaluating bypass techniques and communicating risk levels to government and industry partners, addressing the regulatory uncertainty that initially triggered the export controls.