Google has expanded the capabilities of its AI music generation tool, Lyria, with the launch of Lyria 3 Pro, enabling the creation of songs up to three minutes in length. This builds on last month’s update that allowed the Gemini application to produce 30-second tracks.
The new Pro version offers improved understanding of musical composition, allowing users to request specific elements like intros, verses, choruses, and bridges. Google provided the following example of how the tool can be used:
Lyria 3 Pro is now available to subscribers of Google AI Plus (10 tracks per day), Pro (20 tracks per day), and Ultra (50 tracks per day) within the Gemini application. The company suggests use cases include creating custom soundtracks for video blogs, podcasts, or instructional videos. Users can access the feature by selecting “Create music” from the “Tools” menu.
Beyond the Gemini app, Lyria 3 and Lyria 3 Pro will likewise be integrated into Google Vids, allowing users to add custom music to creative and marketing projects. This functionality will be available to Google Workspace customers and Google AI Pro/Ultra subscribers starting this week.
The technology is also being made available to developers and professionals through several platforms:
- Vertex AI: Lyria 3 Pro is now in public preview on Vertex AI, designed for businesses needing scalable, on-demand audio generation. It enables organizations to expand high-fidelity production, from quickly creating custom game soundtracks to integrating it into creative tools, music, and video platforms.
- Google AI Studio and Gemini API: For developers building the next generation of creative tools, Lyria 3 enhances music awareness and structural consistency to provide creative flexibility. Lyria 3 Pro is now available with Lyria RealTime in AI Studio.
- ProducerAI: Recently launched, ProducerAI is a collaborative music creation tool built by musicians seeking new ways to improve their creative process. With Lyria 3 Pro, ProducerAI offers an agent experience designed to help artists, producers, and songwriters of all skill levels create full tracks. It is available to both free and paid subscribers globally.
Addressing safety and responsible AI practices, Google stated that Lyria 3 “was designed and trained using materials from YouTube and Google that we have the rights to use under our Terms of Service, partner agreements, and applicable law.” The model is also designed to avoid imitating artists.
…if a prompt specifies a creator, the model treats it as broad inspiration
Additional filters are in place to “check outputs against existing content,” and all generations include a SynthID watermark.
FTC: We use affiliate links to earn revenue. More
