Skip to main content

Live models and pricing

Model availability changes, so always choose from the live portal.

Use the live catalog​

The portal publishes model names, availability, and pricing dynamically. This documentation never embeds fixed model IDs, exact prices, or multipliers. Open models and pricing.

Choose a model​

  1. Open the models and pricing page and confirm that the target model is currently available.
  2. Copy the displayed model ID and replace YOUR_MODEL in your configuration.
  3. Send a minimal test request before enabling streaming or tool use.

Beginner selection tips​

  • Choose a model that matches the protocol and client you will use instead of copying a name from an old guide.
  • For the first setup, use an available model with controlled cost to validate the path, then switch to the production model.
  • If the client can fetch a model list, compare it with the live portal. If it cannot, paste the model ID manually.

The model field​

Its location varies by protocol, but every example uses YOUR_MODEL. Do not infer a model name from an old screenshot.

{
  "model": "YOUR_MODEL"
}

Model and protocol capability​

Availability depends on more than the model name: the Key group, channel endpoint, and request protocol must all match.

  • OpenAI compatible: confirm that the model supports Chat Completions, Responses, or the target media endpoint.
  • Claude Messages: confirm that the model is available through the Claude group and use /v1/messages.
  • Gemini generateContent: put the model ID in the URL path and confirm the model supports the Gemini capability.
  • Images, audio, embeddings, reranking, and moderation: the portal must expose the corresponding capability; unsupported requests may return 400 or 404.