Lyria 3 Pro Preview

model·generative music model·active

Google's preview flagship for structured full-length songs.

Recorded facts

Official site	https://ai.google.dev/gemini-api/docs/models/lyria-3-pro-preview ↗
Geography	unknown
model id	lyria-3-pro-preview
api	Gemini Interactions API; Gemini generateContent API
inputs	text; image
license	not publicly disclosed
outputs	44.1 kHz stereo MP3; lyrics; song structure
weights	not publicly disclosed
modalities	44.1 kHz stereo MP3; image; lyrics; song structure; text
dependencies	unknown
model family	Lyria 3
access method	Gemini API; Google AI Studio
exact version	lyria-3-pro-preview
repository links	unknown
evaluation results	result: Lyria 3 improved over Lyria 2; numeric scores not published · source: https://deepmind.google/models/model-cards/lyria-3/ · benchmark: internal expert and automated framework
evidence separation	developer claims: Google describes Lyria 3 as its most capable music generation model to date. · measured results: · independent evaluations:
developer maintainer	Google DeepMind
material us relevance	Publicly accessible in or materially offered to the United States market as of snapshot.
documented capabilities	text-to-music; image-to-music; generation duration: couple of minutes; vocals and instrumentals
products exposing model	Gemini API; Google AI Studio; Google products listed in the model card
commercial use conditions	Subject to Gemini API terms and Google generative-AI policies.
known geographic or policy restrictions	Preview model; Requests for specific artist voices or copyrighted lyrics may be blocked
safety watermarking disclosure provenance	Prompts are safety filtered; Generated audio includes an imperceptible SynthID watermark