On Friday, a landmark event occurred: the U.S. Department of Commerce officially lifted a two-week ban on the export of the advanced artificial intelligence model Claude Mythos 5 by Anthropic. This decision, outlined in a letter from Secretary Howard Lutnick to Anthropic's CTO Tom Brown, grants access to the model for over 100 organizations from Appendix A — a list that includes major U.S. corporations and government agencies. From now on, transferring Mythos 5 to these verified partners no longer requires a special license.

From Glasswing to Global Access

Until now, Mythos was used exclusively within the Glasswing program — a vulnerability discovery project involving approximately 150 organizations from 15 countries. Recall that Mythos previously identified critical errors in classified government systems within hours during testing. Now, the restrictions that blocked the launch of both Mythos and its more powerful "counterpart" Fable 5 have been officially lifted.

It's worth noting that the reason for the block was concerns from Amazon — one of Anthropic's largest investors. The company's researchers warned that Fable 5 could be hacked for potentially dangerous purposes. However, the current decision demonstrates a shift in approach: instead of a total ban, authorities are opting for controlled distribution among trusted partners.

Fable 5 Awaits Its Turn

Unlike Mythos, the fate of Fable 5 remains uncertain. According to sources close to the negotiations, the release process has moved forward, but no specific timeline has been announced. Previously, Fable 5 was the most powerful public AI tool available to all subscribers. Now, its fate is directly tied to new regulations.

On June 2, an executive order was signed that introduces a voluntary channel for federal review of advanced AI models. Developers can now submit their solutions for cybersecurity checks 30 days before launch. This effectively creates a new system of restrictions, reminiscent of controls on AI chip exports to China.

Notably, OpenAI followed the same path on Friday, restricting access to the most powerful version of GPT-5.6, called Sol, to approximately 20 partners who received state approval. Weaker versions, Terra and Luna, became available to the general public.

Global Dependence and Resistance

Countries in Europe and other regions are already expressing concern: they find themselves heavily dependent on Washington's decisions regarding access to new developments. At the same time, leaders of major cybersecurity companies, including former Facebook security chief Alex Stamos, as well as heads of Nvidia, Adobe, and Zoom, are urging authorities to abandon restrictive measures.

My analysis: The decision on Mythos 5 is a compromise between innovation and security that sets a new standard for the entire industry. The question is not whether Fable 5 will be released, but under what conditions. If Anthropic agrees to a government audit before launch, this could set a precedent for all major AI models. Meanwhile, the market is closely watching who gets access to Fable 5 first — this will determine the new balance of power in cybersecurity and analytics.