Skip to content
worth noting New models

Google has launched the Lyria 3.5 music generation model with the ability to edit individual sections of a track

independently confirmed

Google has launched the Lyria 3.5 model in Google Flow Music with improved melodies, lyrics and vocals. The new Selective Section Painting feature allows users to edit a section of a track without regenerating it. Tracks last 30 seconds to 3 minutes; the company has not commented on the training data.

Google has launched a new music generation model, Lyria 3.5, in Google Flow Music. According to the company, it offers richer and more natural-sounding melodic structures, higher-quality lyrics with better adherence to prompts and song structure, and more expressive, emotionally convincing vocals with improved pronunciation. Users also have more precise control over the tempo and output length.

Tracks generated by the model can last from 30 seconds to 3 minutes. A new feature called Selective Section Painting allows users to edit a specific section of a track or expand a short melody into a full song without having to regenerate the entire track. Users can also fine-tune the tempo and duration of individual elements of a track, such as vocals, drums or bass.

Sources vary in how much detail they provide about the training data for the model. When Lyria 3 was first introduced, Google stated that it had trained the model on materials that Google and YouTube had rights to under terms of service, partner agreements and applicable laws, without providing further details. According to The Decoder, Google did not respond to a question about the training data specifically for Lyria 3.5.

What changed

Why it matters

The ability to edit individual sections of a track without having to regenerate the entire track shifts the workflow for AI-generated music from a one-off attempt to iterative editing, which has practical applications for content creators, video production and hobby musicians. At the same time, uncertainty about the training data persists, which is relevant to businesses considering licensing and copyright risks.

Two audiences, two different impacts

What this means

01

For individuals

Users creating music can now edit selected passages of a track or expand a short melody into a full song without having to regenerate the entire track, and set the tempo and duration of vocals, drums or bass more precisely.

What to do Try generating and editing tracks in Google Flow Music using the Selective Section Painting feature.
More practical updates →
02

For a business

Companies creating video content, advertisements or marketing materials can use Google Flow Music with the Lyria 3.5 model to create custom music with more precise control over length and edits to individual sections of a track, but the lack of transparency about the training data poses a risk for commercial use.

Sales and marketing
What to decide When using music from the Lyria 3.5 model commercially, check the licensing terms for Google Flow Music, as the company has not disclosed details about the training data.
More business impacts →
AI creativity Flow Music music generation Google Google DeepMind Google Flow Music Lyria 3.5

Check the original

Event sources

independently confirmed · 2 publishers, 1 independent. We count feeds from the same owner only once.

2
Google DeepMind Blog primary source · first detected We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control The Decoder (daily AI news) independent context Google's Lyria 3.5 music model now lets users edit individual track sections without starting over