Google has launched Lyria 3.5 in Google Flow Music, featuring improvements in musicality, lyrics, vocals, and creative control. This update enhances the AI music generation capabilities of the platform.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
On July 29, 2026, Google DeepMind launched Lyria 3.5, integrated into the Flow Music app, bringing improvements in musicality, lyrics, vocals, and creative control, but without disclosing specific technical metrics or benchmark results.
On July 29, 2026, Google DeepMind announced the release of Lyria 3.5, the latest version of its music generation model, now integrated into the Google Flow Music app. The official blog states that the model brings "significant progress" in musicality, lyrics, vocals, and creative control, but does not provide any quantitative metrics or comparisons with previous versions.
The release of Lyria 3.5 marks Google's continued investment in AI music generation, but the announcement lacks technical details such as model parameter size, training data, or inference speed.
According to the official blog, Lyria 3.5's improvements focus on four areas: musicality, lyrics, vocals, and creative control. In terms of musicality, the model can create "richer, more complex melodic structures" and outputs sound "more natural." Lyrics generation quality has improved, showing "better prompt adherence and structural awareness." For vocals, the model delivers "more realistic, emotionally nuanced" singing, with improved pronunciation. In creative control, users can now "more easily control the speed and duration of outputs."
These descriptions are all qualitative, without any verifiable benchmarks or user study data. For example, there is no explanation of how "more natural" is defined, nor any metrics for evaluating lyrics quality.
Lyria 3.5 is available today in Google Flow Music, an app designed to help users "create the songs you love." The blog emphasizes "creative control" as a core goal, but does not specify the granularity of control or integration with other tools (e.g., text-to-music, melody editing).
This release is Google's latest move in AI music generation, but the announcement does not mention any third-party evaluations, user feedback, or comparisons with competitors (e.g., OpenAI's Jukebox or Meta's MusicGen).
This article is entirely based on the Google DeepMind official blog, a first-party source. All descriptions of model capabilities are official claims, lacking independent verification or quantitative data. Readers should treat it as a product announcement, not peer-reviewed research.
Lyria 3.5 is an incremental update to Google's music generation model, integrated into the Flow Music app, with qualitative improvements in musicality, lyrics, vocals, and creative control, but lacking technical details and independent verification.
Primary report
Primary source
Also covered by 1 sources