US Commerce Department Greenlights Return of Anthropic Fable 5 and Mythos 5 Models After Security Review

The Sudden Shutdown of Frontier Models

On June 9, 2026, Anthropic launched Claude Fable 5 and Claude Mythos 5, the most powerful artificial intelligence models the company had ever publicly released. Just three days later, on June 12, the U.S. Department of Commerce abruptly ordered Anthropic to take both models offline. This directive came from the Bureau of Industry and Security, which imposed strict export controls that required a government license for any foreign nationals, including Anthropic's own foreign employees, to access the models. Consequently, Anthropic had to disable access for all of its global customers to remain compliant with federal regulations.

The swift government action was triggered by a security vulnerability. A researcher at Amazon discovered a jailbreak within Fable 5 that could potentially allow malicious actors to facilitate cyberattacks. Amazon Chief Executive Officer Andy Jassy communicated these findings to Trump administration officials, raising immediate alarms about national security and the risk of the models being diverted to foreign adversaries. Senior administration officials argued that Anthropic's leadership had not initially recognized the severity of the cybersecurity risks associated with the newly released models.

Easing Restrictions on Mythos 5

Following two weeks of intense negotiations and technical reviews, the Commerce Department began easing its restrictions. On June 26, 2026, Commerce Secretary Howard Lutnick sent a letter to Anthropic stating that the company's engagement with the government had yielded significant progress. Lutnick determined that appropriate safeguards had been established to allow trusted partners to access the technology.

Under the revised terms, a government license is no longer required to export, reexport, or transfer the Claude Mythos 5 model to a specific list of approved entities, known as Annex A. This waiver also extends to the foreign national employees of those approved organizations and to Anthropic's own foreign national staff. However, export controls remain strictly in place for any organizations or individuals not explicitly cleared in the administration's letter.

The Full Return of Fable 5

The regulatory resolution was completed on Tuesday, June 30, when Lutnick sent a follow-up letter to Anthropic co-founder Tom Brown. The letter officially cleared Fable 5 to return to public access starting Wednesday, July 1. Lutnick confirmed the decision publicly on X, noting that the government had worked closely with Anthropic to analyze and approve the model to ensure alignment across federal agencies while maintaining American leadership in AI.

To secure the return of Fable 5, Anthropic agreed to several ongoing oversight measures. Under the new agreement, the company, led by CEO Dario Amodei, must continue collaborating with the U.S. government on safety protocols, standards, and release guidelines for Mythos, Fable, and all future models. Furthermore, Anthropic has committed to immediately informing federal officials of any malicious activity detected on its platforms. Anthropic publicly thanked its users for their patience on Tuesday evening and promised to share further updates on the rollout soon.

By utilizing export controls as a rapid-response kill switch for software vulnerabilities, the federal government has established a highly interventionist precedent that leaves AI developers operating under the constant threat of sudden, unilateral shutdowns.