US Government Forces Anthropic to Suspend Fable 5 and Mythos 5 Worldwide

In an unprecedented move, the US government has issued an export control directive under national security authorities ordering Anthropic to suspend all access to its most advanced AI models — **Fable 5** and **Mythos 5** — for any foreign national, both inside and outside the United States. The directive even extends to foreign national Anthropic employees.

As a result, Anthropic has been forced to abruptly disable these models for all customers globally. This marks the first time the US government has directly intervened to shut down access to a commercially deployed frontier AI model after release.

What Happened

On **June 12, 2026 at 5:21pm ET**, Anthropic received a directive from the US government ordering it to suspend access to Fable 5 and Mythos 5 by any foreign national. The letter did not provide specific details of its national security concern, according to Anthropic.

Anthropic published a public statement detailing the situation, explaining that the government believes it has become aware of a method of jailbreaking Fable 5 — a technique to bypass the model's safeguards.

Anthropic's Response

Anthropic strongly disagrees with the government's assessment. The company states that:

• The alleged jailbreak is **narrow and non-universal** — it can only elicit limited information under specific circumstances
• It essentially involves asking the model to read a specific codebase and identify software flaws
• Other publicly available models (including **OpenAI's GPT-5.5**) can perform the same task without requiring any bypass
• Anthropic conducted **thousands of hours** of red-teaming with the US government, UK AISI, and multiple third-party organisations before launch, finding Fable 5's safeguards substantially more effective than any previously deployed model
• No tester has yet found a universal jailbreak for Fable 5

The company wrote: *"If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers."*

What This Means

Fable 5 and Mythos 5 were only launched on **June 9, 2026** — just three days before this directive. These models represent Anthropic's most advanced frontier AI systems, with Fable 5 being their flagship reasoning model and Mythos 5 their specialised variant.

Access to all other Anthropic models (Claude Opus 4.8, Sonnet, Haiku) will not be affected by this directive.

Anthropic has stated it is **working to restore access as soon as possible**, calling the situation a misunderstanding.

Industry Implications

This action raises serious questions about the future of frontier AI development and deployment:

• If narrow, non-universal jailbreaks become grounds for recalling deployed models, every frontier model provider could face similar shutdowns
• The lack of a transparent, statutory process for such decisions undermines the predictability that companies need to invest in AI safety research
• Non-US users of frontier AI models now face the reality of being locked out of state-of-the-art AI systems due to US export controls

The situation echoes earlier debates about AI export controls but takes them to a new level — applying restrictions to already-deployed commercial products rather than pre-release review.

What Happens Next

Anthropic has promised to share more details over the next 24 hours. The company continues to comply with the legal directive while pushing back on the government's reasoning. For customers — particularly enterprise users outside the US — this represents a sudden and disruptive loss of access to models they had integrated into their workflows just days ago.

We will update this story as new information becomes available.

For related reading, check out our analysis of AI export controls and what they mean for self-hosted AI and our guide on running open-source models that can't be taken away.

*Image credit: Anthropic (anthropic.com/news)*