
Lyria 3 Pro Preview
google/lyria-3-pro-preview
Music generation model for full-length songs from text or images with vocals and structure
Capabilities
ReasoningNo
Tool callingNo
Structured outputNo
TemperatureSupported
AttachmentYes
Open weightsNo
Limits
Context131,072 tokens
Max inputN/A
Max output8,192 tokens
Details
Familylyria
Knowledge cutoffN/A
Release date2026-03-25
Last updated2026-03-25
Modalities
Inputtext, image
Outputtext, audio
Comments (0)
Sign in to leave a comment
No comments yet.