Choose a Model

Every request includes a model ID. Copy the exact ID from the live Models page rather than typing a display name from memory.

{
  "model": "MODEL_ID"
}

Model ID versus display name

A model card may show a friendly name alongside the identifier used by the API. Your request and tool configuration must use the model ID.

Model IDs are case-sensitive unless the client explicitly normalizes them. Preserve punctuation, prefixes, and version suffixes.

Match the model to the endpoint

Choose a model that supports the capability you need:

  • Chat or text generation
  • Responses-based agent workflows
  • Embeddings
  • Image generation or editing
  • Speech generation or transcription
  • Video generation
  • Reranking
  • Claude-compatible Messages
  • Gemini-compatible generateContent

A model working with one endpoint does not guarantee that it works with another.

List models through the API

curl https://getrouter.ai/v1/models \
  -H "x-goog-api-key: YOUR_API_KEY"

The response uses an OpenAI-compatible model list. Gemini-compatible clients can use:

curl https://getrouter.ai/v1beta/models \
  -H "x-goog-api-key: YOUR_API_KEY"

When a tool has no model dropdown

Many tools allow manual model entry. Paste the exact model ID from Models. If the tool validates against a cached list, refresh the provider, reload the model list, or restart the tool after changing its Base URL and API key.

Check price before use

Models can use different billing units and prices. Review Pricing before running long contexts, agent loops, large media jobs, or automated tasks.

Warning

Do not copy a model ID from an unrelated provider's documentation unless the same ID appears in GetRouter's model catalog.

Next steps