Models API

GET /v1/models returns the models available to the current API key. The catalog can change with upstream availability, account permissions, and token restrictions, so avoid maintaining a long-lived static “complete model list” in application code.

List models

curl https://api.nbility.ai/v1/models \
  -H "Authorization: Bearer $NBILITY_API_KEY"

Use jq to print IDs only:

curl -s https://api.nbility.ai/v1/models \
  -H "Authorization: Bearer $NBILITY_API_KEY" \
  | jq -r '.data[].id' \
  | sort

The response follows the OpenAI-compatible list format:

{
  "object": "list",
  "data": [
    {"id": "gpt-5.4", "object": "model"},
    {"id": "claude-sonnet-4-6", "object": "model"}
  ]
}

Those IDs illustrate the format only; use the live API response as the source of truth.

Select a model

1. Call /v1/models with the same API key that will make the request.
2. Choose a text, vision, image-generation, audio, embedding, or reranking model for the task.
3. Copy the returned id exactly into the request's model field. Do not invent a provider prefix or alias.
4. In production, allow configuration changes and handle model retirement or permission changes gracefully.

Appearing in the list does not mean a model supports every protocol or parameter. For example, an image model may not accept Chat Completions, and a text model may not support tools. Test against the relevant endpoint documentation.

A model is missing

• Check whether the token has a model or group restriction.
• Verify balance, subscription, or enterprise workspace permissions.
• Refresh the catalog instead of relying on an old client-side cache.
• If the console shows the model but the API does not, contact support with the request ID.