Google Debuts Lyria 3 Pro for Longer AI Music Creation
Google has officially unveiled Lyria 3 Pro, an advanced music generation model designed to significantly extend the creative capabilities of its predecessor. The rollout comes just one month after the initial release of Lyria 3, marking a rapid progression in the company’s generative audio technology.
Extended Capabilities and Creative Control
The primary upgrade in the Pro version is the ability to generate musical tracks lasting up to three minutes, a substantial leap from the 30-second limit imposed on the standard Lyria 3 model. Beyond duration, Google has enhanced the model’s architectural understanding of music.
Users can now exert finer creative control over their compositions by specifying structural components directly within their prompts. The model is engineered to distinguish and arrange distinct elements, including:
- Intros
- Verses
- Choruses
- Bridges
Integration Across Ecosystems
Google is deploying Lyria 3 Pro across both consumer and enterprise platforms. Within the Gemini app, the model is currently being made available exclusively to paid subscribers. Furthermore, the technology is being integrated into Google Vids and ProducerAI, the music production tool acquired by Google last month.
For developers and corporate clients, the model is reaching enterprise tools through several channels:
- Vertex AI (currently in public preview)
- Gemini API
- AI Studio
Training Data and Ethical Safeguards
Addressing concerns regarding copyright and artistic integrity, Google stated that the model was trained using data from its partners alongside permissible content from YouTube and its own internal datasets. The company explicitly stated that the model is not designed to mimic specific artists.
However, when users incorporate an artist’s name into a prompt, the system is programmed to draw “broad inspiration” from that specific style. To ensure transparency, all audio produced by both Lyria 3 and Lyria 3 Pro is embedded with SynthID, a digital watermark indicating the use of artificial intelligence.
This launch arrives as the broader streaming industry intensifies its focus on AI-generated content. Recently, Spotify introduced new features to help artists manage misattributed tracks, while Deezer has deployed identification tools to flag AI-generated music across streaming platforms.
For further details on the technical specifications of the new model, visit the official Google announcement.