Lyria 3.5 in Gemini: How to Make a Full AI Music Track

Google’s Lyria 3.5 brings music generation to the Gemini app. The September 4, 2026 launch focuses on more expressive vocals, richer arrangements and the option to choose shorter or longer tracks. It offers a quick route from an idea to a piece of music, whether the goal is a video backing track, a birthday song or a ringtone.

How do you create a track?

Open Gemini on the web or mobile app and describe the music you want. Google says you can specify a genre and choose a vocal or instrumental style. Templates provide another starting point if you do not know how to phrase a prompt. A useful request names the mood, tempo, instruments and intended use, then asks for an adjustment after hearing the first result.

Can it make longer music?

Yes. Google says users can choose short or longer tracks in this version. The announcement does not make a blanket promise that every request will become a commercial-length song. Choose the length that fits the job and inspect the generated arrangement before committing it to a project.

Where is Lyria 3.5 available?

Google says the model is available to Gemini users globally on the web and mobile app. It also lists Google Flow Music for artists and AI creatives, the Gemini API and Google AI Studio for developers, and Google Vids. The right entry point depends on whether you want a quick personal track, a creative workflow or an application integration.

What makes a better prompt?

• State the purpose: a short intro, a calm background bed or a celebratory piece.
• Name a musical direction: acoustic, electronic, cinematic or another genre.
• Choose vocals or instrumental and specify the desired energy.
• Give feedback about what to change after listening.

This is more useful than an open-ended request for “a song,” because the output needs to match a particular use. Compare several versions for clarity and pacing rather than assuming the first version is final.

Who is this for?

Casual users can experiment in Gemini; creators can test ideas in Flow Music; developers can use the API or AI Studio. Google describes Lyria 3.5 as its best-sounding music generation model at launch, but the practical question is whether the output fits the project you have in mind.

In short: the new model broadens Gemini’s music tools with vocal and instrumental choices, templates and flexible track length, with global app access announced by Google.