Product
Claude Fable 5, now available in Enjamb
Anthropic’s most capable generally available model is now an opt-in choice for demanding, long-horizon work in Enjamb.
by Enjamb Team
Claude Fable 5 is now available in Enjamb as an opt-in model for demanding, long-running work where sustained reasoning and agentic execution matter most.
Anthropic describes Fable 5 as its most capable generally available model, built for ambitious coding, professional work, research, and multi-stage tasks that can run for long periods before review.
Built for the longest horizon
Fable 5 is designed to plan across stages, delegate to sub-agents, check its own work, and continue through tasks that earlier models could not sustain. It supports a one-million-token context window and outputs up to 128,000 tokens.
That makes it a potential fit for unusually large evidence reviews, long technical investigations, complex implementation work, and other tasks where the work itself may span many steps and artifacts.
Opt-in, with provider terms made explicit
Fable 5 does not use the same data-handling terms as every other supported model. Anthropic applies a mandatory 30-day retention period to Fable 5 requests, and the model is not eligible for zero-data-retention arrangements.
Anthropic also applies safety classifiers that may refuse or redirect requests in some higher-risk domains. Enjamb presents Fable 5 as an opt-in choice so teams can evaluate those provider terms before selecting it for a workflow.
Provider availability update: Anthropic restored Fable 5 access on July 1, 2026, after a temporary provider-wide suspension.
Governance remains part of the run
Choosing a more capable model does not bypass Enjamb’s controls. Connector permissions, requester identity, approval gates, activity history, and source provenance still apply around the work.
Fable 5 should be selected only when its longer-horizon capability fits the task and its provider retention and safety terms fit the organization’s requirements.
