The United States government has officially lifted export controls on Anthropic’s advanced artificial intelligence models, Claude Fable 5 and Mythos 5, effective 30 June 2026. This decision follows weeks of intense review after regulators initially imposed the restrictions due to security vulnerabilities identified within the software.
Investigation And Resolution
The export ban was triggered on 13 June 2026, when authorities raised concerns regarding a jailbreak vulnerability discovered in the Fable 5 model. The initial government directive was unconventional, requiring Anthropic to block access to the tools by all foreign nationals, including its own employees, marking a shift in how the U.S. approaches the regulation of commercial AI.
Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America’s leadership in AI.
Commerce Secretary Howard Lutnick confirmed the oversight process, which cleared the way for the broad restoration of access. On 26 June, the department had already signaled its intent to ease these measures by granting a partial exception that allowed vetted institutions to deploy the Mythos 5 model.
Broader Implications For AI Regulation
This development represents a landmark moment in international technology policy, as it is the first time the U.S. government has applied traditional trade enforcement tools to restrict commercial AI software rather than hardware like semiconductors. Industry analysts view the move as a trial balloon for future oversight protocols, suggesting that the government is moving toward a formalized system of security-vetted rollouts for frontier AI technologies.
Anthropic confirmed that global access to the platforms was scheduled to be fully restored as of 1 July 2026. While the immediate restriction is lifted, the collaborative process used to resolve the dispute is expected to serve as a framework for how the government manages national security risks posed by rapid advancements in generative AI.