Why did Anthropic suspend Fable 5 and Mythos 5?

The US government issued an export control directive citing national security authorities. The order required Anthropic to disable access for all foreign nationals, which in practice forced Anthropic to suspend the models for every customer to ensure compliance.

Are other Claude models affected?

No. Anthropic confirmed that access to all other models — including the rest of the Claude lineup — is unaffected. Only Fable 5 and Mythos 5 are suspended.

Was Fable 5 actually jailbroken?

Anthropic says it reviewed a demonstration of a narrow, non-universal jailbreak that surfaced minor, previously known vulnerabilities. The company argues the same outputs are available from other public models such as GPT-5.5 and that no universal jailbreak has been found.

When will Fable 5 and Mythos 5 come back?

Anthropic has not committed to a date. The company is complying with the directive, disagrees with the reasoning, and says it is working to restore access as soon as possible.

What should enterprise customers do right now?

Switch production workflows to other Claude models (Sonnet, Opus, Haiku tiers that remain available) or to comparable models from OpenAI, Google or open-source providers. Re-evaluate model dependencies and keep a multi-provider fallback for any frontier-class workflow.

Anthropic Suspends Fable 5 & Mythos 5 After US Export Directive

The directive in one paragraph

The US government, citing national security authorities, ordered Anthropic to suspend all access to Fable 5 and Mythos 5 by any foreign national — inside or outside the United States — including Anthropic's own foreign-national employees. Anthropic received the letter at 5:21pm ET and concluded that the only way to comply was to shut both models off for everyone. Access to all other Anthropic models is unaffected.

What the government appears to be worried about

The directive did not include specific national-security details. Anthropic's understanding is that the government believes someone discovered a way to "jailbreak" Fable 5. After reviewing a demonstration, Anthropic says the technique surfaced a small number of previously known, minor vulnerabilities — the kind that other public models, including OpenAI's GPT-5.5, also produce without any bypass at all.

Anthropic's defense, summarized

Fable's safeguards were red-teamed for thousands of hours with the US government, the UK AISI, and third parties before launch.
No universal jailbreak — one that broadly bypasses safeguards — has been found.
Anthropic uses a defense-in-depth strategy: narrow jailbreaks are expected, monitoring catches abuse, and 30-day data retention exists specifically to research them.
No disclosed jailbreak has produced a concerning real-world harm; the cited demonstration was "asking the model to read a specific codebase and fix any software flaws" — capability already widely available.

Why this matters beyond Anthropic

If a narrow, non-universal jailpath becomes a sufficient basis to recall a commercial frontier model used by hundreds of millions of people, the precedent affects every frontier provider — OpenAI, Google DeepMind, Meta, xAI, Mistral. Anthropic argues the action does not match the "transparent, fair, clear and grounded in technical facts" standard it has publicly endorsed for government deployment blocks.

What enterprise users should do this week

Move Fable-5 and Mythos-5 production traffic to other Claude tiers or comparable frontier models.
Add a multi-provider fallback to any agent or pipeline that depends on a single frontier model.
Audit data retention and compliance assumptions — they were tuned to Fable's 30-day policy.
Track Anthropic's follow-up disclosures; the company says more details are coming within 24 hours of the original statement.

Bottom line

This is the first time a US national-security directive has forced a frontier-model provider to pull live models from production for every customer at once. Whether or not you use Claude, the policy precedent — and the multi-provider hedging it forces — now applies to the entire AI stack.

Anthropic Suspends Fable 5 and Mythos 5 After US Export Directive