Claude Fable 5 will be available globally again starting Wednesday after the U.S. lifted export controls, with new safeguards in place.
Claude Fable 5 Returns to Global Access
Starting Wednesday, Claude Fable 5 will be accessible globally again, following the lifting of U.S. export controls that had previously forced Anthropic to take it offline. The restrictions had been in place since June 12, limiting access to both Claude Fable 5 and its advanced counterpart, Mythos 5.
In a recent update, Anthropic announced that a new safety classifier has been developed to effectively prevent the jailbreak method that had been discovered, blocking it in over 99% of instances. This classifier will automatically redirect any flagged requests to Claude Opus 4.8, accompanied by a notification to the user.
Export Controls and Model Downtime
The U.S. Department of Commerce's decision to impose export controls stemmed from a jailbreak technique identified by Amazon researchers, which revealed vulnerabilities in the model. To avoid potential compliance issues, Anthropic opted to disable both models for all users globally.
“We’ve been informed that the Department of Commerce has lifted export restrictions on Claude Fable 5 and Mythos 5. We will start restoring access tomorrow,”
— Anthropic (@AnthropicAI) June 30, 2026.
Access Plans and Market Competition
Access to Claude Fable 5 will vary based on subscription plans, with Pro, Max, Team, and select Enterprise users receiving it for up to half of their weekly usage limits until July 7. After this period, users will transition to a system of usage credits.
During the downtime, competitors gained an advantage. OpenAI’s GPT-5.5-Cyber outperformed Mythos 5 on the CyberGym benchmark, scoring 85.6% compared to Mythos 5’s 83.8%. Additionally, other companies, including those from China, launched their own vulnerability detection systems, while alternative smart routing methods using less powerful models matched or exceeded Mythos’s performance.
Future Collaboration and Security Measures
Anthropic is advocating a collaborative approach with Amazon, Microsoft, and Google to address future jailbreak threats, focusing on evaluating techniques based on various criteria such as capability gain and ease of weaponization. They have also committed to providing government access for testing new models prior to their public release and have initiated a HackerOne program for researchers to report bypass techniques for Fable 5’s protections.
