New to this? Read it in simple words
- Google added its Lyria 3.5 music model to Gemini and to developer tools. It opened the model to more people on September 4.
- A user can ask for a full song with words or only music. Through the developer API, Google lists a price of eight US cents per song.
- This makes it cheap to create many first versions. The main work then becomes choosing, editing and checking rights.
- Google says the model has better singing, words and sound quality. No independent test proves it is the best music model.
- API
- A way for developers to use a service directly from their own software.
Google has added its Lyria 3.5 music model to Gemini and to tools for developers. A user can ask for a full song with words or only music. The low price makes music easy to create, but editing and rights still need human care.
What changed in September
Google first showed Lyria 3.5 in its Flow Music product in July. On September 4, it opened the model to more people. It is now available in Gemini on the web and on phones. Developers can also use it through Google AI Studio and the Gemini API.
Users can choose a music style, ask for singing or only instruments, and create short or long tracks. Google says the model has better singing, words, and sound quality. These are Google’s own claims, and there is no independent test that proves it is the best music model.
Google lists a price of eight cents to generate one full song. Lyria 3.5 is available in several Google products.
The price changes creative work
Google lists a price of eight US cents for one full song through the developer API. This makes it cheap to create many first versions. The main work then moves to choosing a good version, changing it, checking rights, and making sure it fits the project.
A text prompt is a simple start, but musicians often need more control. They may want to keep the chorus, change only the middle, or edit separate instruments. Good tools should let people improve one part without making a completely new song.
People need to know where music came from
Google says Lyria music includes SynthID. This is a hidden digital mark that can help identify AI-made audio. It may still work after common changes to the file. However, a hidden mark cannot answer every question about ownership or permission.
Users should know when AI helped make a song. Creators should also keep a clear record of their own choices and edits. Lyria 3.5 makes the first song fast and cheap. Human judgment is still needed before anyone shares or sells it.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1
- 2
- 3
Quality statements are Google’s claims. We found no independent listening test for this version, so we do not rank it.
Gemini can now make a full song. Silicon AI News, September 7, 2026. https://siliconainews.com/stories/lyria-35-gemini-rollout
Google added its Lyria 3.5 music model to Gemini and developer tools, making full songs cheap to create, though rights need human care. How we rate the mood
The mood shows how the news frames the event for the tech and AI world: as an opportunity (6 and above) or as a loss (below 5). It is not our verdict on who is right.