lyria 3
Google music model that turns context into songsText, image, or video becomes a 30-second track
Highlights
What you can do with it
- Video creators generate background music for YouTube Shorts
- Marketing teams make unique music for brand campaigns
- Music lovers craft personal 30-second tracks
How to use it
- 1 Open the Gemini app, where Lyria 3 is available first on the web
- 2 Choose text, image, or video input to describe the music
- 3 Refine style, tone, speed, mood, and language with prompts
- 4 Generate a 30-second segment, listen, and adjust as needed
- 5 Save the track for personal or promotional use
Who it is for
Pros
- Free version available with plan limits
- Multimodal input from text, images, and video
- Automatic lyrics and clearer song structure
- SynthID watermark supports AI audio transparency
Limitations
- Output is limited to 30-second segments
- Free version has plan limits; higher tiers cost more
- Lyria 3 is consumer-facing, while enterprise use follows Lyria 2's API
- Model access and terms may change over time
What Lyria 3 is
Lyria 3 is the newest model in Google DeepMind's Lyria family, launched in Gemini on February 18, 2026. It turns text, image, and video context into 30-second music segments. It can generate lyrics automatically from a prompt, produce more coherent song structures, and deliver higher fidelity in short-format output. It supports rich multimodal control, so creators can guide music with the visual context of images and videos as well as with words. Music generated in Gemini carries an imperceptible SynthID audio watermark that helps verify the source of AI-generated or edited audio. There is a free version with plan limits, and paid tiers such as Plus, Pro, and Ultra raise generation limits.
What you can do with it
- Turn a text, image, or video prompt into a 30-second music segment.
- Let the model write lyrics automatically from your prompt.
- Guide style, tone, speed, mood, and language with multimodal input.
-Save tracks for personal projects or brand promotion.
Who it is for
- Video creators who need quick background music for Shorts and social.
-Marketing teams that want unique music for campaigns.
-Everyday users who want to make music without production skills.
What to watch out for
Output is limited to 30-second segments, so it is not built for full-length songs. The free version has plan limits, and higher generation limits require paid tiers. Lyria 3 is consumer-facing, while enterprise integration follows the Lyria 2 API path, so plan accordingly. Because it lives inside Gemini, access and features may change over time, and the SynthID watermark means AI origin can be identified, which is worth noting if you plan to present the audio as your own recording.
FAQ
What is Lyria 3?
It is Google DeepMind's newest Lyria model in Gemini that turns text, image, and video context into 30-second music segments.
Where can I use it?
It is available through the Gemini application, first on the web version, with the mobile version to follow.
Is there a free option?
There is a free version with plan limits, and paid tiers like Plus, Pro, and Ultra offer higher generation limits.
Why it is free
What the official page says： free version
Free limits, trial length and features are decided by the official site and can change at any time.
Product preview
Screenshot archived from the listing page, for preview only.
