Model Discovery
Find models by region in OCI Generative AI.
Use model discovery to find models in a specified region and review details such as:
- Model identifiers and vendors.
- Capabilities, including chat, embedding, rerank, and fine-tuning.
- Supported input and output types, for example, text, images, audio, or video.
- Regional availability and support for on-demand or dedicated serving.
- Whether a model supports OCI OpenAI-Compatible endpoints.
- Model parameters that you can set.
Without filters, results include all pretrained and fine-tuned models across all regions.
Filter the Results
Use filters to narrow the search by:
- One or more realms or regions.
- A specific model.
- Model capability, serving mode, or support for OCI OpenAI-Compatible endpoints.
- Whether a model is deprecated or retired for on-demand or dedicated serving.
- Model access: hosted or proxy.
Hosted models run in OCI Generative AI. Proxy models are hosted externally by their providers. When you call a proxy model through the service in an OCI region, the request is sent to a provider location. For more information, see Notes for External Calls.
For example, find chat models that support on-demand serving in your region, or avoid models retiring soon for that serving mode.
Use the ListModelDiscovery operation to find models and view their details.
Use the list-model-discovery command to find models and view their details.
Model discovery isn't available in the Console.