What is lyria 3
Lyria 3 is the latest iteration in the Lyria family of AI music generators, launched on February 18, 2026, within Google's Gemini. It is designed to transform text, image, and video prompts into 30-second music clips. Key advancements include auto-generated lyrics, improved song structure, and integrated SynthID watermarking for provenance.
How to use lyria 3
Lyria 3 is primarily accessed through the Gemini app. Users can input text, image, or video prompts to guide music generation. The model automatically handles aspects like lyrics and song structure (intro, verse, chorus, etc.). For enterprise use, Lyria 2 is currently available via Vertex AI.
Features of lyria 3
- Multimodal Prompting: Accepts text, image, and video inputs to influence music generation.
- Auto-Generated Lyrics: Can create lyrics automatically based on prompt intent.
- Enhanced Song Structure: Produces more coherent musical pieces with distinct sections like intros, verses, and choruses.
- High-Fidelity Short-Form Output: Generates cleaner, more layered 30-second music clips.
- SynthID Watermarking: Embeds imperceptible watermarks for AI-generated audio verification.
- Multilingual Support: Available in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese.
Use Cases of lyria 3
- Creating soundtracks for short-form content like YouTube Shorts.
- Assisting non-musicians in generating music from creative prompts.
- Rapid content creation workflows requiring quick music generation.
- Experimentation with AI-driven music composition using visual or textual cues.
Pricing
Lyria 3 is available through Gemini with a free tier that has usage limits. Higher generation limits are available through Gemini Plus, Pro, and Ultra subscription tiers.
FAQ
0. Is "lyria3" the same as "Lyria 3"? Yes. "lyria3" is a common search term, while "Lyria 3" is the official name.
1. What is the biggest practical difference between Lyria 3 and Lyria 2? Lyria 3 offers multimodal prompting and auto-lyrics within Gemini, while Lyria 2 is the current API-accessible model for enterprise integration via Vertex AI.
2. Can I call Lyria 3 directly from Vertex AI today?
No, public Lyria 3 API access is not confirmed yet. Vertex AI currently offers Lyria 2 (lyria-002).
3. Which model should teams use if they need API reliability now? Teams should use Lyria 2 on Vertex AI for current enterprise workflows.
4. How long are Lyria 3 tracks in Gemini right now? The current output length is approximately 30 seconds per generation.
5. Does Lyria 3 generate lyrics automatically? Yes, Lyria 3 can generate lyrics automatically from prompt intent.
6. Can prompts use images or video context? Yes, Lyria 3 supports multimodal steering from text and visual context.
7. Is there watermarking for generated audio? Yes, Gemini-generated Lyria audio includes SynthID watermarking for provenance.
8. Can I ask for exact imitation of a specific artist? Lyria 3 includes safeguards that treat artist references as broad inspiration rather than direct cloning.
9. Is Lyria 3 intended to replace professional music production? Its current positioning is for fast expressive creation and soundtrack support, particularly for short-form content, not as a full replacement for professional pipelines.
10. Who can access Lyria 3 now? Access is described for users aged 18+ through Gemini, with a web-first rollout and tier-dependent usage limits.
