US Government Lifts AI Export Controls on Anthropic Fable 5 and Mythos 5 Models

On June 30, 2026, the Trump administration officially ended a weekslong standoff with artificial intelligence developer Anthropic by lifting strict export controls on its most powerful models, Claude Fable 5 and Mythos 5. The Department of Commerce informed the company of the decision, prompting Anthropic to announce that it would begin restoring user access the following day. The resolution marks the end of a high-stakes dispute over national security and regulatory oversight that had temporarily frozen the deployment of Anthropic flagship technology.

The Origins of the Export Ban

The conflict began earlier in June 2026, shortly after Anthropic publicly released Claude Fable 5. Although the company built safeguards into Fable 5 to prevent its misuse in cyberattacks, a trusted partner, later identified as Amazon, discovered a jailbreak. This vulnerability allowed users to bypass the model safety guardrails to access restricted capabilities. Anthropic initially argued that the vulnerability was relatively simple and common among other publicly available models, asserting that recalling a commercial model over a narrow jailbreak could halt the entire frontier AI industry.

However, the Commerce Department responded by issuing a broad export control order. The directive required Anthropic to block access to the model by foreign nationals, a restriction so sweeping that it applied to Anthropic own foreign employees. Because of the order's reach, the company was forced to completely disable customer access to both Fable 5 and Mythos 5, the latter being a more powerful version designed with fewer guardrails for specialized testing by key partners to secure critical software.

Diplomatic Shifts and New Safeguards

To resolve the freeze, Anthropic shifted its strategy and communication style with federal officials. In negotiations with Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross, Anthropic CEO Dario Amodei was replaced in meetings by co-founder Tom Brown, whom government officials reportedly preferred on a personal level. Instead of continuing to debate whether jailbreaks could be entirely prevented, Anthropic focused on addressing the administration's immediate concerns.

The company developed and implemented a new safeguard designed to block the specific security vulnerabilities reported by Amazon. Anthropic also agreed to proactively identify and address future security risks while working closely with federal authorities on release protocols and safety standards for Mythos, Fable, and subsequent AI models. These concessions paved the way for a gradual easing of restrictions, which began last week when the government allowed limited access to Mythos 5 for approved companies and agencies.

Restoring Access and Moving Forward

The complete lifting of the export controls was finalized in a letter from Lutnick to Brown. In the letter, Lutnick confirmed that a license is no longer required for the export, reexport, or in-country transfer of either the Mythos or Fable models, including deemed exports. Lutnick publicly noted that his department worked closely with Anthropic over a two-week period to analyze and approve Fable 5, aiming to align federal oversight and maintain American leadership in artificial intelligence.

With the regulatory hurdle cleared, Anthropic is set to restore public access to Fable 5 and resume its testing of Mythos 5. This rapid pivot from a severe export ban to a collaborative agreement suggests that the future of frontier AI development will depend heavily on a company's willingness to grant federal regulators direct influence over its safety engineering and internal leadership dynamics.