Choose the model for your work

50 media models across image, video and audio.

A model can support several operations. Speech variants within one shared model interface are listed in its settings and counted together.

All 50 models

Create images

Turn a prompt into an image. Choose a model, aspect ratio and supported resolution before spending credits.

Edit images

Use an existing image as the starting point. Describe the change and choose a model that supports your reference requirements.

Create video

Build a shot from a prompt, image or supported reference media. Compare duration, resolution and reference limits before generating.

Edit video

Transform an existing video using supported editing models. Source duration, resolution and reference support differ by model.

Design voices

Design a reusable voice, remix an eligible voice or create a clone from permitted samples. Preview candidates before saving a voice.

Text to speech

Give a script a voice. Choose a compatible saved voice and speech model, then review the generated audio.

Create sound effects

Describe a sound and choose its supported duration. Compare short effects and longer audio models using the same cost assumptions.

Create music

Create a musical bed or a longer track. Choose a model by supported duration, vocal controls and credit cost.

Language models

The Responses API also has 16 language model contracts. Product-surface access is separate from media generation.