Google DeepMind Releases Lyria 3: An Advanced Music Generation AI Model that Turns Photos and Text into Custom Tracks with Included Lyrics and Vocals

2026-02-18 · Source: Machine Learning ML & Generative AI News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Content Creation & Production · Depth: Intermediate, quick

Summary

Google DeepMind has released Lyria 3, a new multimodal generative AI model integrated into the Gemini app, capable of converting text prompts and photos into 30-second music tracks. This advanced model generates full arrangements, including vocals and lyrics, achieving superior long-range coherence and 48kHz audio quality. Designed for both creators and engineers, Lyria 3 incorporates SynthID, an inaudible digital watermarking technology, to ensure the detectability of AI-generated content even after extensive editing. This release, alongside the Music AI Sandbox, aims to elevate generative audio from basic MIDI loops to professional-grade, "human-in-the-loop" synthesis, establishing a new benchmark for the AI music industry in 2026.

Key takeaway

For AI Product Managers evaluating new creative tools, Lyria 3 represents a significant advancement in generative audio, offering high-fidelity output with integrated vocals and lyrics. You should consider its multimodal input capabilities and the inclusion of SynthID for content provenance when assessing its potential for new applications or integrations within your product roadmap.

Key insights

Lyria 3 is a multimodal AI generating high-fidelity music with vocals and lyrics from text and photos.

Principles

• Multimodal input enhances creative AI output.
• Digital watermarking ensures AI content traceability.

Method

Lyria 3 converts text prompts and photos into 30-second music tracks with vocals and lyrics, utilizing SynthID for watermarking.

In practice

• Generate custom music for content creation.
• Create tracks with specific lyrical themes.

Topics

• Music Generation AI
• Multimodal AI
• Generative Audio
• AI Watermarking
• Google DeepMind

Best for: Machine Learning Engineer, AI Scientist, AI Product Manager, AI Engineer, Research Scientist, Creative Technologist

Related on AIssential

• Build with Lyria 3, our newest music generation model — Lyria 3 offers AI music generation with structural coherence, expressive vocals, and granular control via text and multimodal inputs.
• Google launches AI music generator Lyria 3 Pro, says it was trained on data it has the right to use — Google's Lyria 3 Pro advances AI music generation with improved structural understanding and broad platform integration.
• Google adds AI music generation to Gemini with Lyria 3 — Google's Lyria 3 model enables AI music generation in Gemini and YouTube, featuring mood-matching and content identification.
• Record scratch—Google's Lyria 3 AI music model is coming to Gemini today — Google's Gemini app now offers AI music generation with Lyria 3, featuring SynthID watermarking and copyright considerations.
• Lyria 3 Pro: Create longer tracks in more Google products — Lyria 3 Pro extends AI music generation to 3-minute tracks with structural control across Google's ecosystem.
• Google brings AI music to the masses — Google's Lyria 3 model integration into Gemini democratizes AI music creation for a massive consumer audience.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning ML & Generative AI News.