U.S. Government Orders Anthropic to Shut Down Its Most Powerful AI Models:
Claude Fable 5 and Claude Mythos 5 pulled offline worldwide amid national security concerns — and a public dispute over what really triggered the order:
A Sudden Shutdown Order:
On Friday, the U.S. government ordered Anthropic to immediately disable access to two of its most advanced AI models, Claude Fable 5 and Claude Mythos 5, citing national security concerns. Anthropic confirmed on X that it has complied with the directive.
According to Anthropic, the order was received Friday at 5:21 p.m. ET and applies to all users worldwide, not just the foreign nationals the export control order was nominally targeting. Anthropic's other models remain unaffected.
Models Affected: Claude Fable 5 and Claude Mythos 5
Order Received: Friday, 5:21 PM ET
Scope: Worldwide, all users
Why Mythos Was Already Locked Down:
Claude Mythos 5 is Anthropic's most capable model to date, previewed in early April and kept under tight restriction ever since. Anthropic says the model demonstrated an exceptional ability to find security vulnerabilities, reportedly identifying flaws in every major operating system and web browser it was tested against.
Rather than release it broadly, Anthropic launched Project Glasswing, a controlled program sharing Mythos with roughly 50 vetted organizations — including Amazon, Apple, Google, Microsoft, and CrowdStrike — for defensive cybersecurity research.
Fable 5: A Guardrailed Commercial Release:
Released just three days before the shutdown, Claude Fable 5 was positioned as Anthropic's commercial answer to Mythos — the same underlying capability, but fitted with guardrails blocking high-risk responses in areas like cybersecurity and biology. According to benchmark tests from Vals AI, Fable 5 was immediately the most capable AI model available to the public.
"We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people." — Anthropic
The Hidden AI War
Nobody Is Telling You About
Our latest documentary deep-dive into the geopolitical struggle for machine intelligence dominance. Explore the two paths of AI development: open source vs. closed architecture.
The Real Trigger: A Disputed Jailbreak Claim:
While the government's directive is officially framed as an export control measure, Anthropic says its understanding is that the actual concern centers on a claimed jailbreak of Fable 5. The company states the government has so far provided only verbal evidence of what it describes as a "potential narrow, non-universal jailbreak."
Anthropic characterizes the jailbreak as essentially prompting the model to analyze a specific codebase and identify software flaws — a capability the company says is already widely available in other public models, including OpenAI's GPT-5.5, and one routinely used by cybersecurity professionals for defensive purposes. The company also points to its layered safety architecture, arguing that independent classifier systems operating separately from the model itself provide a backstop — meaning that even if a user pushes Fable past a refusal, protections against the most dangerous outputs remain active.
Anthropic Pushes Back Hard:
In a lengthy blog post, Anthropic made its frustration clear, warning that if this standard of recalling a deployed commercial model over a narrow jailbreak finding were applied industry-wide, it would essentially halt new model deployments across all frontier AI providers.
An Ironic Turn for the 'Safety-First' Lab:
The timing is notable: Anthropic is widely expected to pursue an IPO this year and has built much of its public identity around being the safety-conscious alternative to other AI labs. Observers have pointed out the irony that the same caution Anthropic displayed in restricting Mythos — promoted as a model so capable it couldn't be released publicly — appears to have drawn exactly the kind of regulatory scrutiny now disrupting its business.
OpenAI CEO Sam Altman has weighed in too. Back in April, he told podcaster Ashlee Vance that Anthropic's handling of Mythos amounted to "fear-based marketing," comparing it to building a bomb, threatening to drop it, and then selling a bomb shelter. While Altman didn't predict a government shutdown specifically, his broader point — that consistently framing your AI as uniquely dangerous tends to get noticed by regulators — appears to have landed.
What This Means for Enterprise AI Buyers:
This episode is a reminder that the frontier AI landscape can shift overnight due to factors entirely outside a business's control — regulatory action, export restrictions, or sudden model deprecations. For companies building critical workflows on top of any single frontier model, that kind of dependency carries real operational risk.
At Otherworlds AI, our Agent+ Business AI platform is designed with this reality in mind, giving businesses flexible, model-agnostic enterprise AI infrastructure that isn't locked to any single provider's roadmap or regulatory exposure.
Combined with Google Opal automated workflows, our approach helps organizations build resilient AI operations that can adapt as the frontier model landscape continues to evolve — because as this week's news shows, it can evolve fast.
