First celebrated, then shut down: Anthropic’s most powerful AI model to date disappeared just days after its launch—at the behest of the U.S. government. Here’s what’s behind it and what it means for you.
What is Claude Fable 5?
A few days ago, Anthropic released Claude Fable 5, its most powerful publicly available model to date. It belongs to the new “Mythos” class, which ranks above Claude Opus, and is designed for particularly complex, long-running tasks—such as coding sessions that independently plan, test, and refine code over the course of hours.
What happened?
Since June 12, Fable 5 has abruptly disappeared—from Claude Chat, Claude Code, and all other interfaces. The reason is unusual: Citing national security concerns, the U.S. government issued an export control directive prohibiting access by foreign nationals. As a result, Anthropic had to disable the model for all customers worldwide.
The Background
Apparently, the government assumes that Fable 5 can be “jailbroken”—that is, deliberately circumvented. Anthropic strongly disagrees: The vulnerability in question is narrow in scope and harmless, and comparable capabilities have long been freely available in competing models such as GPT-5.5. The company openly refers to this as a misunderstanding and considers a recall to be disproportionate.
Impact on Claude Code users
The good news is that all other models—Opus 4.8, Sonnet, and Haiku—will continue to run without any restrictions. Anyone who has been using Fable 5 for major refactorings or long autonomous sessions should switch to Opus 4.8 for the time being. For most development tasks, day-to-day operations will remain unchanged.
| Conclusion: This case illustrates how quickly government intervention can shut down a commercial AI model—and raises the question of how far such intervention might go in the future. |
What happens next?
Anthropic has announced that it will restore access “as soon as possible,” but has not provided a specific timeline. Regardless of the outcome, the case is likely to spark a broader debate: about the balance between security concerns and the availability of powerful AI for businesses and self-employed individuals.