Details
- Google DeepMind introduced Lyria 3 Pro, an advanced AI music generation model enabling tracks up to 3 minutes long with structural awareness for elements like intros, verses, choruses, and bridges, building on last month's Lyria 3 limited to 30-second clips.
- Involves Google DeepMind, Vertex AI, Google AI Studio, Gemini API, Google Vids, Gemini app, and ProducerAI; collaborations with artists like Yung Spielburg and François K via Music AI Sandbox.
- Lyria 3 Pro offers enhanced customization, high-fidelity 48kHz stereo audio from text or images, vocals, lyrics, and complex transitions, integrated for developers, businesses, and creators.
- Expands from Lyria 3's beta in Gemini app (Feb 2026) to broader rollouts including public preview on Vertex AI, AI Studio, Vids for Workspace/AI subscribers, and global ProducerAI access.
- Positions Google against Suno and Udio in generative music; features SynthID watermarking, responsible training on licensed YouTube/Google data, and filters to avoid mimicking artists.
Impact
Google's Lyria 3 Pro escalates the AI music race by matching Suno/Udio's length and structure via massive distribution in Gemini, Vertex AI, and Workspace, accelerating enterprise adoption for gaming, video, and content tools. Over 12-24 months, it could shift R&D toward agentic music agents, boost developer ecosystems building on Gemini API, and pressure incumbents amid rising scrutiny on AI training data and IP safeguards.
