Anthropic Restores AI Models Fable 5 and Mythos 5 After U.S. Export Control Lift

Anthropic has announced the restoration of its AI models, Claude Fable 5 and Claude Mythos 5, after the U.S. government lifted export controls that had been imposed on June 12, 2026. The model Fable 5 resumed global access on July 1, 2026, while Mythos 5, a variant with fewer restrictions, is set to be accessed by designated U.S. organizations approved after a review on June 26, 2026.

The export controls were enacted due to a cybersecurity incident reported by Amazon, where a jailbreak was found that could exploit Fable 5 by allowing it to generate harmful code. As a precaution, Anthropic suspended access to both models for all users, fearing violations of the new regulations due to the inability to verify users' nationalities immediately.

This incident emphasizes the growing complexity of AI governance and oversight. Anthropic has committed to enhancing collaboration with U.S. agencies moving forward. This includes sharing threat intelligence and offering agencies early access to new models and their safety protocols before public release. Anthropic is also working alongside tech giants like Amazon, Microsoft, and Google to develop a standardized framework for evaluating the risks associated with AI software vulnerabilities, which are known as jailbreaks.

As part of this initiative, Anthropic's efforts tie into an executive order from early June 2026, focusing on enhancing AI security. Consequently, models like Fable 5 and Mythos 5 will likely undergo increased scrutiny and regulation in the future, raising questions about the balance between innovation in AI and government oversight.

The impact of the initial export suspension was reflected in financial markets, where a perpetual futures contract for Anthropic listed on the Hyperliquid exchange dropped about 3.7% as concerns grew regarding the timing of the company's anticipated public listing.