Google AI Developers
@googleaidevs
Create studio-quality, high-fidelity compositions with Lyria 3 and Lyria 3 Pro, our newest music generation model. Now available in public preview via the Gemini API and in @GoogleAIStudio 🎶
6retweets
View on XGoogle Releases Lyria 3 Music Generation Models via Gemini API
· Updated
Lyria 3 and Lyria 3 Pro, Google DeepMind's newest music generation models, are now in public preview via the Gemini API.
lyria-3-pro-preview generates full-length tracks up to approximately three minutes; lyria-3-clip-preview produces 30-second clips optimized for speed and high-volume requests. Both support realistic vocals, multiple languages, and genres from pop to funk to Motown.What sets Lyria 3 apart is its compositional control: tempo conditioning with high accuracy, time-aligned lyrics with start and end points within a track, and image-to-music input that shapes the audio's mood and atmosphere. Every generated track also carries a SynthID watermark that verifies it as Google AI-generated even after modification.
Google is also launching a music playground in Google AI Studio — requiring a paid API key — with text mode and composer mode for building songs section by section. For developers building audio-enabled apps, Lyria 3 is worth evaluating as an integration layer.
