The US government shut down Anthropic’s most powerful model in 72 hours. But 200 organizations still have the raw version. A two-tier AI access regime is now official policy.
What Happened
Fable 5 launched June 9. By June 12, the US government ordered Anthropic to shut it down — a jailbreak had been discovered that bypassed safety guardrails. Anthropic complied immediately, cutting off all public and commercial access.
But Bloomberg reports that 200 Project Glasswing organizations — including Cisco and Dragos — still have access to the raw, unguardrailed Mythos Preview for cyber defense. The guardrailed version is shut down. The unguardrailed version runs freely for cleared organizations.
Meanwhile, Trump told G7 leaders that Anthropic talks are “going fine” — the same company whose model his government just shut down is also in active negotiations with the White House.
The key insight: Frontier AI is becoming a controlled substance. Anthropic is becoming its licensed distributor. The model that gets shut down for the public but keeps running for the Pentagon is not losing — it is winning the game that matters.
What to Watch Next Week
1. Whether Fable 5 comes back online — and under what conditions
2. How OpenAI and Meta respond if the same standard applies to GPT-5 and Llama
3. Anthropic-White House negotiation outcome
Business Engineer Framework
The Harness Society — The Permission Layer Is Real
Read The Harness Society →








