Google is rolling out Lyria 3.5, its next-generation artificial intelligence music generation model, designed to elevate musicality, lyrics, and vocal expression within its Flow Music platform. According to company announcements, the updated model focuses heavily on structural awareness, prompt adherence, and granular creative controls that let creators manipulate individual elements like tempo, vocals, drums, and bass.
The release marks another step in Google’s rapid iteration of its audio generation technology. Following earlier rollouts of the Lyria architecture earlier in the year—which brought short-form music creation and style controls to tools like YouTube Shorts and Google Vids—Lyria 3.5 introduces deeper creative flexibility and more natural-sounding audio for digital creators.
Enhanced Musicality and Prompt Adherence
At the core of the Lyria 3.5 update are architectural refinements aimed at improving how the AI interprets complex text prompts. Google stated that the model features enhanced prompt adherence, reducing instances where generated lyrics drift away from the user’s initial instructions or theme.
Alongside lyrical consistency, the company noted that Lyria 3.5 delivers more expressive vocals paired with updated pronunciation systems. The model’s musicality engine has been retrained to produce intricate, multi-layered sounds that maintain a more organic feel. Users can continue to prompt the system using descriptive vibes, moods, or scenarios, retaining the flexible style-control features introduced in previous iterations.
Granular Creative Control and Track Lengths
Creative control represents a primary upgrade in the Lyria 3.5 release. Creators utilizing Flow Music gain the ability to adjust specific audio components within an AI-generated piece, altering vocals, basslines, and drum tracks independently.
Furthermore, the update refines tempo management and track duration tools. While earlier versions of the technology primarily focused on generating brief 30-second snippets, Lyria 3.5 expands user capabilities to support track lengths ranging from 30 seconds up to three minutes.
Integration Across Google Ecosystems
Earlier variants of the music generation technology found homes in developer previews and creative suites, including Dream Track integrations for YouTube Shorts and specialized soundtrack generation features within Google Vids.
Users can access the new features directly through Flow Music.
What are your thoughts on AI-generated music and tools like Lyria 3.5? Join the conversation by sharing your perspective in the comments below.
Worth a look