Gemini model resources
| Method | Path | Purpose |
|---|---|---|
| GET | /v1beta/models | List Gemini-protocol model resources |
| GET | /v1beta/models/{model} | Retrieve one resource |
Use https://api.fenno.ai and x-goog-api-key: <FENNO_API_KEY>.
List request
curl https://api.fenno.ai/v1beta/models -H "x-goog-api-key: $FENNO_API_KEY"The response contains models[]. When provided, a resource can include name, displayName, token limits and supportedGenerationMethods. Treat missing capability fields as unknown rather than inventing defaults.
If a paginated response returns nextPageToken, use supported pageToken and pageSize parameters to continue. Do not assume either pagination or its maximum page size is identical across routes.
Resource names
A name can look like models/<MODEL_ID>. When creating a raw URL, append that complete name after /v1beta/, or remove the single models/ prefix before inserting it into /v1beta/models/{model}. Do not produce /models/models/....
The Google GenAI SDK accepts a model ID or a supported models/... resource name. Use the exact available identifier rather than adding another provider’s model-family prefix.
Availability and capabilities
Discovery is a starting point, not proof that every optional feature, tool or modality will work. Verify generateContent and any required streaming or vision behavior with a small request under the intended credentials.
See Generate Content and Gemini CLI setup.