Back to News
New featureFeaturedJuly 1, 2026

Claude Fable 5 is back

The US lifted the export controls and Anthropic restored global access. Fable 5 is again the frontier model on Enterprise plans, with a new safety classifier.

aimodelosanthropicfable-5
Claude Fable 5 is back

On June 30 the US Department of Commerce lifted the export controls on Claude Fable 5 and Mythos 5, and Anthropic restored global access on July 1.

What had happened

On June 12, a US government directive forced Anthropic to suspend Fable 5 for all customers over a jailbreak risk flagged by researchers. Our multi-model architecture did its job: service was never interrupted — frontier tasks automatically fell back to Opus 4.8 (we covered it here).

What changed to bring it back

Anthropic trained a new safety classifier that blocks the reported technique in over 99% of cases. If a Fable 5 request is blocked, it's automatically rerouted to Opus 4.8 — transparent to the user. Commerce's Center for AI Standards and Innovation validated the safeguards.

What it means on Napsix

  • Fable 5 is again the frontier model on Enterprise plans: premium chat, vision, OCR and the most demanding code generation route back to Fable 5 (Mythos-class: 1M context, 128K output, state-of-the-art coding).
  • Opus 4.8 stays as the automatic fallback — the resilience we already proved.
  • It's back in the model picker as "Claude Fable 5 (latest)".

Coverage: Axios · The Verge · BBC.

XIA
public