Models: browse and choose
Filter the marketplace by modality, read billing types and context length, and copy model IDs.
The model marketplace (the "Models" page in the console) is the entry point for choosing and wiring up models: browse every model aggregated by the platform, inspect capabilities and billing, and copy a model ID straight into your requests.
Filter by modality
The left navigation (or the filters at the top of the Models page) groups models by modality:
| Modality | Typical uses |
|---|---|
| Text | Chat, writing, code, translation — plain-text tasks |
| Image | Text-to-image / image understanding |
| Audio | Speech recognition, speech synthesis |
| Video | Video understanding / generation |
| File | Document parsing, long-document Q&A |
What to look at when choosing
- Model ID: the unique identifier for the
modelparameter in requests — copy it from the card rather than typing it; - Billing type: each card shows the model's billing type and unit price; usage is metered in tokens (see Billing basics);
- Context length: how much conversation and document text fits into one request;
- Modal capabilities: which modalities the model accepts as input / produces as output (e.g. "text in, image out").
Wiring up a model
- Confirm the target model is available in the marketplace and copy its model ID;
- Swap
modelin the first request example for that ID; - Check the "Usage" page to confirm the call succeeded and the consumption matches expectations.
The model lineup evolves with platform operations, so keep model IDs in one place in your configuration (an environment variable or config file) to make switching painless.