Anthropic broke through the blockade: Mythos 5 received the "green light" for the top 100 US organizations, with Fable 5 next in line.
The U.S. regulator has lifted the two-week ban on exporting Anthropic's flagship model, Claude Mythos 5. This decision opens access to the advanced AI tool for over 100 leading U.S. corporations and government agencies included in the special Appendix A list. A separate license is no longer required to transfer the model—compliance with verified partner criteria is sufficient.
The decision was formalized by Commerce Secretary Howard Lutnick in a Friday letter to Anthropic's Chief Technology Officer Tom Brown. Essentially, this marks the official end of the standoff between the administration and the company: export restrictions that previously blocked the launch of both Mythos and Fable have been annulled. However, while Mythos is returning to service, the fate of Fable 5 for the general public remains uncertain.
Mythos 5: From Secret Testing to Mass Adoption
Until now, Mythos was used exclusively within the Glasswing program—an initiative for vulnerability discovery involving approximately 150 organizations from 15 countries. During government testing, the model identified critical errors in classified systems within hours, drawing the attention of authorities. Now, the tool, proven effective in cybersecurity, becomes available to key players in the U.S. market.
Notably, investor concerns—particularly from Amazon, Anthropic's largest shareholder—had previously slowed the process. Company researchers warned that Fable 5 could potentially be hacked for dangerous purposes. However, after the signing of an executive order on June 2, which introduces a voluntary federal review channel for advanced AI models, the situation changed. Developers can now submit their solutions for cybersecurity checks 30 days before launch.
Fable 5 and the New Control Regime
The release of Fable 5 has been delayed, but exact timelines have not been announced. Unlike Mythos, access to this model was previously open to all subscribers, and it was once considered the most powerful public AI tool. However, now that Washington is tightening chip exports to China and introducing similar requirements for access to the models themselves, Fable 5 may face new restrictions.
Notably, OpenAI took the same path on the same Friday: the company restricted access to the most powerful version of GPT-5.6, named Sol, to approximately 20 partners approved by the state, while the weaker versions Terra and Luna became available to the general public.
The key trigger for the blocks was the fear that Chinese entities could acquire the technology. Attention was focused on South Korean operator SK Telecom, which was added to the Glasswing list in early June and then had its access restricted. SK Telecom denies any connection to China, but the precedent itself showed how sensitive the issue of distributing advanced AI models has become.
Cryptalist Analysis: Lifting the ban on Mythos 5 is not just a victory for Anthropic, but a signal of the formation of a new architecture for global AI control. The U.S. is creating a precedent where access to the most powerful models will be strictly regulated by the state, not just market mechanisms. Fable 5 will likely receive similar approval, but with even stricter conditions. Europe and other regions have already realized they are completely dependent on Washington's decisions. This marks the beginning of a new era of "digital sovereignty" in AI, where access to technology will become as strategic a resource as chips.