Anthropic has abruptly cut off access to two of its most advanced AI models. On June 12, the company said the US government issued an export control directive to suspend all access to Fable 5 and Mythos 5. As a result, Anthropic disabled both models for every customer worldwide. However, all of its other models, including the main Claude chatbot, remain available.
The directive landed fast. Anthropic says it received the order at 5:21 p.m. ET that same day. Officially, the government cited national security authorities and export control law. Notably, the letter did not spell out the specific concern. The order targets foreign nationals, including Anthropic’s own foreign-national employees. Because the company cannot cleanly separate those users, it chose to pull the models for everyone.
What Fable 5 and Mythos 5 Do
Both models are recent arrivals. Mythos 5 is a cybersecurity-focused model, offered only to partners in Anthropic’s Project Glasswing program. Fable 5, meanwhile, launched on June 9 to bring those Mythos capabilities to the public. In other words, the suspension hits Anthropic’s newest cyber tools rather than its everyday assistant.
The Jailbreak At The Center
The dispute centers on a single jailbreak. According to Anthropic, the government believes it found a way to bypass Fable 5’s safeguards. The technique is narrow, though. Essentially, it involves asking the model to read a specific codebase and fix any software flaws. Anthropic reviewed a demonstration and found only a few minor, previously known vulnerabilities. Furthermore, the company says other public models can find the same flaws without any bypass.
Anthropic also pushed back on the capability claim. It argues the demonstrated skill is widely available elsewhere, including in OpenAI’s GPT-5.5. In fact, the company notes that defenders use the same techniques every day to keep systems safe. It also points to its defense-in-depth approach, which adds 30-day data retention to detect and shut down attacks quickly.
Anthropic Complies Under Protest
Still, Anthropic is following the order. “We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users,” the company wrote. At the same time, it made its objection clear.
We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.
Anthropic
The company also warns of a broader chilling effect. If regulators applied this standard across the industry, it says, new model launches could effectively grind to a halt. Anthropic still supports the government’s right to block unsafe deployments. However, it wants that power used through a process that is “transparent, fair, clear, and grounded in technical facts” — principles it says this action did not follow. For now, the company calls the move a misunderstanding and says it is working to restore access, with more details promised within 24 hours.
