The US administration has lifted the two-week ban on the export of the advanced AI model Claude Mythos 5 from Anthropic. This decision opens access to the technology for more than 100 American organizations, including major corporations and government agencies. The restrictions, which previously blocked both Mythos 5 and its more powerful version Fable 5, are now officially revoked.
US Secretary of Commerce Howard Lutnick confirmed that transferring the model to organizations on a special list (Appendix A) no longer requires a separate license. This resulted from intensive negotiations between Anthropic and the administration, during which key company employees personally visited Washington.
Why this matters for the market
Recall that initially, both models — Mythos and Fable — were disabled due to concerns from Anthropic's largest investor, Amazon, regarding their safety. Researchers warned that Fable 5 could be hacked for potentially dangerous purposes. Until now, Mythos 5 was used exclusively within the Glasswing project — a vulnerability discovery program involving about 150 organizations from 15 countries. Within hours of testing, the model was already identifying errors in classified systems.
The situation is now changing dramatically. Lifting the ban on Mythos 5 is not just a technical decision but a signal of a new AI control system being formed. On June 2, an executive order was signed creating a voluntary channel for federal review of advanced models. Developers can submit their solutions for cybersecurity checks 30 days before launch.
What about Fable 5 and the global context
While Mythos 5 returns, the fate of Fable 5 remains uncertain. Unlike its "big brother," access to Fable was previously open to all subscribers, and for a time it was considered the most powerful public AI tool. However, its release is now postponed, with no timeline announced.
Interestingly, OpenAI followed the same path, restricting access to its most powerful version, GPT-5.6, named Sol, to about 20 partners who received government approval. Weaker versions, Terra and Luna, became available to the general public.
Initially, the blocks were driven by fears that the technology would reach China. Particular attention was focused on South Korean operator SK Telecom, which was temporarily added to the Glasswing list and then had its access restricted. SK Telecom denied any connection to China. Cybersecurity leaders, including former Facebook security chief Alex Stamos, as well as representatives from Nvidia, Adobe, and Zoom, urged authorities to abandon restrictive measures.
My analysis: We are witnessing the birth of a new paradigm — "managed access" to the most powerful AI. Lifting the ban on Mythos 5 for trusted partners is a compromise between innovation and security. However, for the global market, this is an alarming signal: countries in Europe and other regions suddenly find themselves heavily dependent on Washington's decisions. Whether Fable 5 will receive similar approvals will become clear in the coming days, but the trend toward segmenting AI access is already evident.