Anthropic broke the blockade: Claude Mythos 5 received the "green light" for 100 US organizations, Fable 5 is next.
The U.S. administration has lifted a two-week ban on the export of the most powerful artificial intelligence model, Claude Mythos 5, from Anthropic. This decision opens access to cutting-edge technology for over 100 American organizations, including major corporations and government agencies. At the same time, the public launch of the Fable 5 model — an even more advanced product — remains uncertain.
Commerce Secretary Howard Lutnick sent a letter on Friday to Anthropic's Chief Technology Officer Tom Brown, confirming that transferring Mythos 5 to organizations on the "Appendix A" list no longer requires a special license. Current safeguards are deemed sufficient to ensure security.
Why Mythos 5 was blocked and what has changed
The conflict escalated after Anthropic's key investor, Amazon, expressed serious concerns about the safety of the new models. Company researchers warned that Fable 5 could be hacked for potentially dangerous purposes. As a result, both models were temporarily taken offline for external access.
Until then, Mythos 5 was used exclusively within the secret Glasswing project — a vulnerability search program involving approximately 150 organizations from more than 15 countries. Notably, this model discovered critical errors in classified government systems within hours during testing.
Fable 5 awaits — the rules of the game are changing
The fate of Fable 5 remains uncertain. Unlike Mythos, access to this model was previously open to all subscribers, and for a time it was considered the most powerful public AI tool in the world. Now its release is frozen, with no timeline provided.
The situation is compounded by new regulations. On June 2, an executive order was signed creating a voluntary federal review channel for advanced AI models. Developers can submit their solutions for cybersecurity checks 30 days before launch. This effectively establishes a new control system that tightens not only the export of chips to China but also access to the models themselves.
A new reality: geopolitics and AI
Notably, OpenAI has followed the same path, restricting access to the most powerful version of GPT-5.6, called Sol, to approximately 20 government partners. Weaker versions, Terra and Luna, have been made available to the general public.
Initially, the blocks were driven by fears that the technology would reach China. South Korean operator SK Telecom came under scrutiny: it was first added to the Glasswing program, then had its access restricted. The company itself denies any connection to China.
Leaders of major cybersecurity companies, including former Facebook security chief Alex Stamos, as well as executives from Nvidia, Adobe, and Zoom, have publicly called on authorities to abandon restrictive measures. However, as practice shows, Washington is only tightening control.
My analysis: The unblocking of Mythos 5 is not so much a victory for Anthropic as a signal to the market: the U.S. is shaping a new AI security architecture where access to advanced models will be strictly regulated by the state. Fable 5 will likely receive approval, but under much stricter conditions than Mythos. The industry will have to adapt to a new reality where "open" AI is a temporary exception, not the rule.