GatewayProviders

Browse the model catalog

Find an available model and choose how its provider costs are billed.

The model catalog is the current inventory of models available through the Respan Gateway. Use it to find a model identifier, check whether a model is active or deprecated, and understand whether it uses Respan credits or your own provider credentials.

Open Models

Go to the Models page in the Respan platform.

Respan Models catalog showing Active and Deprecated tabs, model search and filters, custom-model actions, model identifiers, status, and Credits or BYOK billing labels.
Browse active or deprecated models, search the catalog, and filter by provider, model type, or status.

Find a model

  1. Select Active for models currently offered or Deprecated to review models you should migrate away from.
  2. Search for a model name, or open filters to narrow by Provider, Model type, or Status.
  3. Copy the model identifier exactly as displayed, including its provider prefix when present.
  4. Pass that identifier as the model in your Gateway request.
1response = client.chat.completions.create(
2    model="openai/gpt-5.6-sol",
3    messages=[{"role": "user", "content": "Hello"}],
4)

The live catalog is the source of truth for current availability. Avoid hard-coding a catalog size in application guidance because models and statuses change over time.

Understand billing labels

LabelMeaningBefore calling the model
CreditsProvider usage is charged against Respan credits.Confirm that the project has sufficient credits.
BYOKThe request uses provider credentials that you connect.Add the required key on the Provider keys page.

Availability and billing mode are separate concerns. An Active model can still require BYOK credentials, and a configured provider key does not make a Deprecated model active.

Add a model that is not listed

The create menu provides two paths:

  • New custom model adds an alias or model configuration for a provider model you can already call.
  • New model request asks Respan to consider a model that is not available in the catalog.

For request-time routing, provider pinning, credential overrides, and custom-model setup, see Providers & models.

Treat capability, cost, and performance metadata as selection signals, then validate the model with your own prompts and evaluations. A catalog entry does not guarantee that a model is suitable for your workload.