Skip to main content

Models Hub

The Models Hub is the catalog of AI models published on the platform. Each model supports two usage modes:

  • Public Endpoint: billed per token, shared endpoint, subject to a rate limit
  • Private Deployment: billed by resource usage hours, dedicated endpoint

Browse and filter​

You can filter models by type, or search by model name, provider, and description. Each model card shows the model type, name, provider, description, and tags; featured models are specially marked.

Model details​

Click a model card to view:

  • Pricing: public endpoints are billed per token and show the rate limit; billing varies by model type. Private endpoints are billed by the GPU/CPU resource hours used in the deployment.
  • Model Information: the model name and last update time.

From the details page you can also:

  • Playground: open the interactive trial page. You need an API key before trying a model. Note that trials are also billed.
  • API Endpoint: open the "Setup and chat as code" panel and choose Python, JavaScript, or cURL to get code samples and the API path. Replace your-api-key-here in the sample with your API key before use.
  • Private Endpoint: open the deployment panel, set the endpoint name, description, and resource specification, then deploy. See Private Endpoints.