Anthropic has broken through the blockade: Mythos 5 is now open to the top 100 US organizations, with Fable 5 on the horizon
Washington has officially lifted export restrictions on Anthropic's Claude Mythos 5 model. This decision, announced on Friday, opens access to one of the most powerful AI systems to over a hundred American organizations, including leading corporations and government agencies.
This move puts an end to a two-week standoff between the Trump administration and Anthropic. As a reminder, both models — Mythos and Fable — were previously disabled after one of Anthropic's key investors, Amazon, expressed serious concerns about their safety. The company's researchers warned that Fable 5 could be hacked for potentially dangerous purposes.
Mythos 5: From Glasswing to Full Access
Until now, Mythos was used exclusively within the Glasswing project — a vulnerability discovery program involving about 150 organizations from more than 15 countries. This model has already proven itself, identifying errors in classified systems within hours during government testing. Now, as stated by Commerce Secretary Howard Lutnick, existing safeguards allow trusted partners from Appendix A to receive the model without additional licensing.
Fable 5: Awaiting the Green Light
The fate of Fable 5, previously considered the most powerful public AI tool, remains uncertain. Its release process has moved forward, but exact timelines are not yet disclosed. Unlike Mythos, access to Fable 5 was previously open to all subscribers, making it particularly vulnerable from a regulatory standpoint.
The situation is compounded by the formation of a new restriction system. On June 2, an executive order was signed, opening a voluntary channel for federal review of advanced AI models. Developers can now submit their solutions for cybersecurity checks 30 days before launch. Simultaneously, Washington is tightening exports of AI chips to China, and these requirements now extend to access to the models themselves.
Notably, OpenAI followed the same path on Friday, restricting access to the most powerful version of GPT-5.6, called Sol, to about 20 state-approved partners. The weaker versions, Terra and Luna, became available to the general public.
Initially, the reason for the blocks was concern that the technology would reach China. The focus was on South Korean operator SK Telecom, which was added to the Glasswing list in early June and then had its access restricted. SK Telecom denied any connection to China.
Many leaders of major cybersecurity companies, including former Facebook security chief Alex Stamos, as well as leaders from Nvidia, Adobe, and Zoom, urged authorities to abandon restrictive measures. European countries and other regions are already noting that they have become heavily dependent on Washington's decisions regarding access to new developments.
My comment: This precedent marks a new stage in the global AI race. The US is shaping a model of "controlled distribution," where access to advanced technologies becomes a privilege, not a commodity. For the crypto industry, which is increasingly integrating AI for blockchain analysis and DeFi protocols, this means that access to the most powerful tools may be limited not only by technological but also by political barriers. Keep an eye on developments with Fable 5 — it will serve as the litmus test for the new regulatory regime.