← All PostsClaude Fable 5 Pulled by US Gov Three Days After Launch
AI modelsAnthropicsecurity

Claude Fable 5 Pulled by US Gov Three Days After Launch

June 13, 2026

Fable 5 had three days. That's how long it took for the US government to pull the plug on Anthropic's most capable model — not because it was unsafe, but because of an alleged jailbreak that Anthropic says is narrow, non-universal, and already reproducible with models already in the wild.

Fable 5 launched June 9 and was globally suspended on June 12–13 after a US export control directive. As of right now, it's gone from every model picker across Claude.ai, GitHub Copilot, AWS Bedrock, and the API.

What the Government Said

The US government, citing national security authorities, issued an export control directive on June 12 at 5:21 PM ET 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 letter did not provide specific details of its national security concern. What the government did share — verbally — was a report of a potential jailbreak technique involving asking the model to read a specific codebase and identify software flaws.

What Anthropic Actually Found

Anthropic pushed back. Hard.

Anthropic reviewed a demonstration of the specific jailbreak technique and found the level of capability displayed is widely available from other models, including OpenAI's GPT-5.5, and is used every day by defenders who keep systems safe.

Anthropic said it has not even received a disclosure of a concerning non-universal potential jailbreak that led to a harmful result. The potential jailbreaks disclosed to them are either entirely benign responses or minor findings that provide no Mythos-specific uplift.

Their position: this is a narrow jailbreak, not a universal one. And no one has demonstrated a universal bypass.

Why They Killed It Anyway

Because they had to. The net effect of the order is that Anthropic must abruptly disable Fable 5 and Mythos 5 for all customers to ensure compliance, since the directive covers any foreign national anywhere — making it impossible to selectively restrict access without a full shutdown.

GitHub Copilot followed immediately: access to Claude Fable 5 was suspended across all Copilot experiences effective June 12, with all other Claude models — Opus 4.8, Sonnet 4.6, Haiku 4.5 — remaining available and unaffected.

AWS similarly revoked access to both Fable 5 and Mythos 5 on Bedrock for all users, citing compliance with the US government export control directive.

The Data Retention Catch

Worth knowing: Fable 5 required 30-day data retention to operate Anthropic's safety classifiers — a policy change that Anthropic acknowledged carries real costs for customers, but enables them to research and mitigate jailbreaks quickly.

That's part of why Project Glasswing exists — the framework Anthropic built specifically around monitoring Mythos-class model deployments for zero-day safety issues.

Anthropic's Position

Anthropic stated they disagree that a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, they believe it would essentially halt all new model deployments for all frontier model providers.

The company says it is complying with the government's legal directive but believes this is a misunderstanding and is working to restore access as soon as possible. No return date has been given.

In the meantime, Fable 5 remains fully available on the Claude API and consumption-based Enterprise plans — suspended access is what's in effect now, but Anthropic's staging plan had already included a billing shift on June 23 where subscription plan users would have needed usage credits anyway.

Opus 4.8 is your best available option right now. Not the same, but it's there.

Source: Anthropic Statement, The New Stack, Claude Status