Use the live catalog for production-ready SENDFU-operated inference services. Separately documented SENDFU-developed variants show their own access and deployment state.
Only model services returned by the live catalog are currently accepted for inference. Each retains its third-party base-model identity while the customer-facing inference service is operated by SENDFU in SG.
Specifications are source-labeled references. Prices are the versioned SENDFU Official Rate Card in USD per token; the request-received UTC timestamp selects and freezes any scheduled rate window.
This classification is reserved for a separately identified fine-tuned variant with SENDFU-controlled artifacts—not a renamed third-party base model.
A documented model variant for mature-audience fictional dialogue, roleplay, and long-form narrative work, subject to adult eligibility and controlled review.
View full model cardControlled Access. Not listed in the public API catalog and not currently enabled on the production inference route.
Review the access processModel origin, service operation, and current access are separate facts. The page keeps those boundaries visible.
Third-party base models retain their published identifiers and developer attribution. SENDFU-developed is used only for a separately controlled fine-tuned variant.
For models in the live catalog, SENDFU operates the customer-facing inference service in SG. This does not transfer ownership of third-party model weights.
“Available” in GET /v1/models plus a valid SENDFU Bearer API key is the current public API access boundary. Controlled documentation is not API entitlement.