On June 12, 2026, something unprecedented happened in the history of artificial intelligence: the United States government ordered a private company to shut down its most advanced AI models for every user on Earth.
The targets were Claude Fable 5 and Mythos 5 — Anthropic's two most powerful frontier models by a catastrophic margin. Fable 5 was a general-purpose reasoning powerhouse, while Mythos 5 was purpose-built for cybersecurity, capable of scanning codebases and identifying zero-day vulnerabilities at inhuman speed.
Just three days after their public launch, both went dark.
The Trigger
The U.S. Department of Commerce, citing national security authorities, issued an export-control directive that stunned the AI world. The order required Anthropic to "suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees."
The reported trigger was an alleged jailbreak that exposed a code-analysis capability powerful enough to worry national security agencies. The Pentagon and NSA reportedly flagged the models' potential to be weaponized against critical American infrastructure — ironic, given that Mythos 5's entire purpose was to defend it.
Anthropic had no choice. At 5:21 p.m. ET on June 12, both models were disabled globally.
The Fallout
The ban sent shockwaves through the enterprise AI world. Companies that had built workflows around Fable 5's reasoning capabilities scrambled for alternatives. The cybersecurity community lost access to Mythos 5's vulnerability-hunting abilities — tools that Cloudflare had recently tested and described as capable of finding zero-days that teams of human security engineers had missed.
Project Glasswing, Anthropic's initiative to deploy Mythos-class models to critical infrastructure operators, ground to a halt. The project had already expanded to operators in 15 countries across power, water, healthcare, and communications sectors. Now, only US-based organizations retained access.
The tension was palpable: the same government that had long pressured the industry to build stronger AI safety measures had now identified safety itself as a national security threat.
The Unfolding Resolution
After 15 days of diplomatic back-and-forth, the landscape shifted. On June 27, Axios reported the Trump administration was close to allowing Fable 5's return. By June 29, Mythos 5 access had been partially restored for a limited group of American organizations under Project Glasswing, though Fable 5 remained offline pending sign-off from the Pentagon and NSA.
Secretary of Commerce Howard Lutnick negotiated new protocols with Anthropic — compliance frameworks, usage monitoring, and ongoing government oversight. The terms effectively establish a new paradigm: frontier AI models operating under continuous export control, treated more like nuclear technology than software.
What It Means
This isn't just a story about one company's models. It's a precedent. The Fable 5 / Mythos 5 ban marks the first time a government has forcibly removed an AI model from global availability. The Export Administration Regulations (EAR) were applied to AI in a way that creates a playbook for future actions.
The implications are staggering. If the US can unilaterally shut down frontier models for all foreign nationals, what does global AI development look like? Other nations will accelerate their own domestic AI efforts. US AI companies face a choice between building the most capable models possible or keeping them deployable at global scale. You may not be able to do both.
For now, the world watches as Fable 5's fate hangs in the balance — the most powerful general-purpose AI ever built, held hostage by the very government that wants its creators to lead the world.
Photo by Brett Jordan on Unsplash
