Indicates the model should return audio.
Indicates the model should return images.
The modality is unspecified.
Indicates the model should return text
Indicates the model should return video.