Anthropic Restores Claude Fable 5 After US Lifts Export Ban

Anthropic is bringing its Claude Fable 5 and Mythos 5 models back online after the US government lifted a brief export ban.



Fable 5 will be available globally starting Wednesday, July 1, across the Claude platform and partner services. On Claude, Fable 5 will be included for up to 50% of weekly usage limits for Pro, Max, Team, and select Enterprise subscribers through July 7, after which it will be available via usage credits. Anthropic said support for AWS, Google Cloud, and Microsoft Foundry will return as quickly as possible. It has also restored access to Mythos 5 for select US organizations participating in Project Glasswing, an initiative Apple recently joined to help secure critical software infrastructure. The reinstatement follows the mid-June suspension triggered by export restrictions over the models' cybersecurity capabilities.


The export control directive followed a report from Amazon researchers describing a way to bypass Fable 5's safeguards. Before the restrictions were imposed, Amazon CEO Andy Jassy personally alerted the White House about the findings. Anthropic said its own testing found that numerous other AI models could perform the same tasks and that the reported technique did not expose Mythos' unique cybersecurity capabilities. It has since deployed an updated safety classifier that blocks the reported technique in more than 99 percent of cases. Requests that trigger the classifier are now automatically routed to the older Claude Opus 4.8 model.

For Apple developers using Xcode 27's new AI coding features, the change restores access to Anthropic's latest coding models. Apple integrated Anthropic's technology directly into Xcode 27 earlier this month to power autonomous coding agents, and Fable 5 was designed specifically for advanced software engineering workloads.

Moving forward, Anthropic is partnering with Amazon, Google, Microsoft, and other organizations to establish a standardized framework for assessing AI jailbreaks. The proposed system would evaluate jailbreaks based on capability gain, breadth of capability gain, ease of weaponization, and discoverability, helping AI companies and regulators assess new findings using a common standard. Anthropic also agreed to deepen its collaboration with the US government, promising pre-release access for advanced models, rapid information sharing when significant misuse patterns are identified, and expanded joint AI security research.

Please download the iClarified app or follow iClarified on X, Facebook, YouTube, and RSS for more updates.