Claude Fable 5 Returns After Jailbreak Triggered a 19-Day Export Ban

The ban and the bypass

Anthropic shut down Claude Fable 5 and Mythos 5 for all users on June 12 after the US government issued an export control directive covering both models. The order required Anthropic to cut off access for any foreign national, including its own non-citizen employees, and took effect immediately, leaving the company no reliable way to verify user nationality in real time.

The trigger was a claude jailbreak reported by Amazon researchers: a prompting technique that got Fable 5 to identify software vulnerabilities and, in one case, produce code demonstrating how one could be exploited. Anthropic said the same vulnerabilities were identifiable by weaker models, including Opus 4.8 and rivals like GPT-5.5 and Kimi K2.7, and that the bypass did not expose unique anthropic mythos level offensive capabilities.

Back online, with a stronger classifier

Fable 5 becomes available globally starting July 1 across the Claude Platform, Claude.ai, Claude Code, and Claude Cowork. Pro, Max, Team, and select Enterprise plans get it included for up to 50% of weekly usage limits through July 7, after which it shifts to usage credits. Access on AWS, Google Cloud, and Microsoft Foundry will follow.

Mythos 5, the less-restricted sibling built on the same underlying model, is only reopening to roughly 100 vetted US organizations that defend critical infrastructure, under Project Glasswing. Anthropic said it trained a new safety classifier that blocks the reported bypass technique in over 99% of cases; flagged requests now route to Opus 4.8 instead.

Why US builders should watch what comes next

The episode pushed Anthropic toward a formal jailbreak severity framework built with Amazon, Microsoft, and Google, scoring capability gain, breadth, ease of weaponization, and discoverability. It is also opening a HackerOne program for researchers to report anthropic cybersecurity bypasses in Fable 5, and has agreed to give US government partners earlier pre-release access to future frontier ai model launches.

For US enterprise teams, the near-term risk is not model capability but availability. A single national security directive suspended a frontier ai model worldwide overnight, with no advance notice. Operators building production workflows on any single lab's models now have a concrete case for fallback plans.

Sources