Google DeepMind launches Lyria 3.5 music model in Flow Music
Google DeepMind announced the launch of Lyria 3.5, described as its newest music generation model, live today in Google Flow Music. The company says the model delivers significant advancements across musicality, lyrics, and vocal quality, intended to help users craft richer tracks with more creative control.
DeepMind's announcement lists four areas of improvement. On musicality, the model is said to create richer, more complex melodic structures that sound more natural. On lyrics, it is said to generate higher quality output with improved prompt adherence and structural awareness. On vocals, DeepMind says the model brings more expression and emotional nuance to songs, along with improved pronunciation, producing more realistic singing. On creative control, the company says users can more easily set the tempo and duration of generated tracks.
The announcement does not include benchmark scores, listening-test data, or comparisons against the previous Lyria version or rival AI music tools. It does not name individuals or teams behind the release, does not describe the model's architecture or training, and does not mention pricing, a free tier, or regional availability. The post's call to action simply invites users to try Lyria 3.5 today in Flow Music.
Key facts
- Google DeepMind launched Lyria 3.5, its newest music generation model, inside Google Flow Music.
- DeepMind says the model produces richer, more complex melodic structures that sound more natural.
- Lyrics generation is described as higher quality, with improved prompt adherence and structural awareness.
- Vocals are said to carry more expression and emotional nuance, plus improved pronunciation.
- Users get easier control over the tempo and duration of generated tracks.
Why it matters
Lyria is Google DeepMind's music generation line, and Lyria 3.5 is the version the company is now rolling into its Flow Music product. DeepMind frames the release as a broad upgrade touching melody, lyrics, vocal delivery, and the controls creators use to shape output, rather than a single isolated feature. That positions it as a move in the fast-moving AI music generation space, though the announcement offers no benchmarks or direct comparisons to size up how large the claimed gains actually are.
Who it affects
The update targets people who use Google Flow Music to generate songs: musicians, hobbyists, and anyone experimenting with AI-assisted songwriting. DeepMind's announcement does not name specific teams, spokespeople, or partners involved in building or shipping the release.
How to use it
Google says Lyria 3.5 is available now inside Google Flow Music, with a direct invitation to try it today in the product. The announcement gives no pricing, no mention of a free versus paid tier, and no word on regional availability or platform restrictions.
How solid is it
The claims come from a single DeepMind blog post and are self-reported, described in comparative terms such as more natural, higher quality, and more realistic, without accompanying data, listening tests, or third-party evaluation. No benchmark scores, audio samples, or technical detail on the model's architecture or training accompany the announcement.
Risks and caveats
Because the source is a company announcement rather than an independent review, the improvement claims are unverified outside DeepMind's own description. The absence of pricing, availability, and technical detail means prospective users cannot judge from the announcement alone how the update compares to the previous Lyria version or to competing AI music generation tools.
“Our newest music generation model, Lyria 3.5, delivers significant advancements across musicality, lyrics, and vocal quality, empowering you to craft richer tracks.”
— Google DeepMind