Anthropic's Fable 5 and Mythos 5 shut down after White House export order

On June 9th, Anthropic launched two new AI models: Fable 5, touted as the most capable generally available model ever, and Claude Mythos 5, which shares the same underlying architecture but without some safeguards. Just three days later, the White House issued an export control directive demanding Anthropic suspend access to both models for any foreign national, including its own employees. The order came after Amazon and White House conversations about researchers finding a jailbreak that could make Fable 5 serve information useful for cyberattacks. Anthropic responded by shutting down access to both models completely for all customers, stating they disagree with the recall rationale but must comply with the legal directive.

Behind the scenes, additional reports from Semafor suggest the decision was partly driven by fears that a group linked to China had already accessed Mythos 5, raising national security risks of reverse engineering through model distillation. The Trump administration’s swift and sweeping action surprised the industry, as Mythos was already limited to select government agencies and companies, and Fable 5 had additional safeguards. Anthropic spent the weekend in Washington trying to change the president's mind, warning that the recall could alter the trajectory of American AI dominance. The unprecedented shutdown gives new ammunition to global competitors arguing they cannot rely on US technology for critical infrastructure.