The Export Control Precedent: Anthropic Disables Claude Fable 5 and Mythos 5

The U.S. government has effectively halted the global deployment of Anthropic’s most advanced AI models. Following a national security directive, the company disabled Claude Fable 5 and Claude Mythos 5 for every customer worldwide after an export control order was issued.

The directive, sent by Commerce Secretary Howard Lutnick to Anthropic CEO Dario Amodei, targets Mythos-class models. Because the order bars access by any foreign national—including Anthropic’s own employees on U.S. soil—the company determined that selective compliance was impossible. The models were pulled globally at 5:21 pm ET, three days after their launch.

The catalyst for this intervention appears to be a specific security vulnerability. The government has provided verbal evidence of a narrow, non-universal jailbreak involving a technique where the model reads a codebase to identify software flaws. Anthropic notes that other public models, such as OpenAI's GPT-5.5, perform this task without a bypass.

This development signals a shift in how the U.S. government manages frontier AI capabilities. By requiring a license for the export, re-export, or domestic transfer of these models, the administration is treating high-level reasoning capabilities as sensitive dual-use technology.

The implications for the AI industry are significant. Anthropic has argued that if recalling models deployed to hundreds of millions of users over a single narrow vulnerability becomes the industry standard, it will halt frontier model launches across the sector. The company is currently complying with the order while attempting to restore access.

The central question for the industry is whether the era of rapid, global deployment for frontier models is over, replaced by a regime of strict, person-based access controls.