Model catalog

Find a model, check its price and providers, and add models that aren't listed.

Gateway › Models lists every model you can call through the gateway, with its price, context window, and capabilities. Use it to pick the model ID you send as model.

Find a model

Open the catalog

In the platform , go to Gateway › Models. Each row is one model. Input and Output show its cheapest price across providers, in USD per 1M tokens.

The Models list with search, the Active, Provider, Model type, Modality, and Capabilities filters, and columns for input and output price, context, latency, capabilities, and release date

Narrow the list

Type in Search for models… and press Enter.
Filter by Provider, Model type (supported or custom), Modality, and Capabilities.
The list starts on Active models. Switch that filter to Deprecated to find models to migrate off.
Click More filters for context length, input and output pricing, and model age.
Click a column header to sort. Star a model to keep it at the top of the list in this browser.

Click a model. The header shows its maker, release date, lowest price, context window, and capabilities. The Providers tab lists every provider that serves it, each with its own model ID, status, price, and billing label.

The GPT-5.5 detail with maker, release date, price, context window, and capability tags, and a Providers tab listing OpenAI (gpt-5.5), Azure OpenAI (azure/gpt-5.5), and OpenRouter (openrouter/openai/gpt-5.5) with status, input and output cost, and billing

Hover a provider and click the copy icon to copy its model ID. The other tabs hold a Code sample, a Playground, Benchmarks, and Throughput, Latency, and Uptime charts.

Call it

Send the model ID as model. The prefix picks the provider: gpt-5.5 goes to OpenAI and azure/gpt-5.5 goes to Azure OpenAI. client is the OpenAI client from the Gateway quickstart.

Python


response = client.chat.completions.create(
    model="azure/gpt-5.5",
    messages=[{"role": "user", "content": "Hello!"}],
)

Understand billing labels

The Billing column of the Providers tab shows how each provider is paid for. To see it in the list too, click the gear icon (Column options) and turn on Billing.

Label	Who pays
Credits	Your Respan credits.
BYOK + Credits	Your Respan credits, or your own provider key if you’ve added one.
BYOK	Your own provider key only. Credits can’t pay for it.

Billing is per provider, so the same model can be BYOK at one provider and BYOK + Credits at another. When your own key pays, the provider bills your account.

To add a key, click Manage provider keys at the top of the model detail. See Provider keys.
To add credits, see Billing and credits.

Pin a provider

The model ID’s prefix picks the provider. When you can’t change the model ID, for example because a CLI agent always sends claude-sonnet-4-5-20250929, send the X-Respan-Route-Provider header with the prefix of the provider you want, as its model ID shows in the Providers tab. This request goes to Vertex AI:

Terminal


curl https://api.respan.ai/api/chat/completions \
  -H "Authorization: Bearer $RESPAN_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Respan-Route-Provider: vertex_ai" \
  -d '{"model": "claude-sonnet-4-5-20250929", "messages": [{"role": "user", "content": "Hello!"}]}'

To fall back to another model when one fails, see Retries and fallback. To let Respan pick the model for each request, use span-router.

Add a model that isn’t listed

Add a custom model to call a model your provider offers before it’s in the catalog.

Start a custom model

On Gateway › Models, click New custom model.

Name it

Pick the Provider and enter the Model name exactly as the provider names it. Below the name, the dialog shows the Model ID you’ll call. The name can’t match a model the provider already has in the catalog. Click Continue.

The New custom model dialog, step 1 of 4, with Provider set to OpenAI, Model name gpt-5.5-2026-09-15, and the Model ID openai/gpt-5.5-2026-09-15

Set capabilities and pricing

Turn on Function calling and Image input if the model supports them, pick its Supported parameters, and click Continue. Then enter Input cost and Output cost in USD per 1M tokens. The cache costs fill in from the input cost, and you can change them. Rate limit and Max context window are optional. Click Continue.

Respan stores these as the model’s prices. Over the API, send the same cost fields to Create or update a custom model and read the stored prices back from pricing in Get a custom model, which gives each price with its unit and any long-prompt tiers.

Save

Check the summary and click Save. The model appears in the list with a Custom tag. Send its model ID as model.

If no provider you use offers the model, click Request a model to ask for it on the Respan Discord.

Use your own endpoint

To call a self-hosted or other OpenAI-compatible endpoint, add it as a custom provider first. Bring your own key needs the Team plan, unless your organization already has an active provider key.