Google is advancing its foray into generative music technology with the announcement of Lyria 3, a new AI music generation model that will be accessible via its Gemini app. This feature, currently in beta, is expected to roll out to all users over the coming days, provided they are 18 years or older. Users will initially be able to generate songs in a variety of languages, including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with additional languages anticipated in future updates. Free subscribers will find their music creations limited to 30 seconds, while those subscribed to Google AI Plus, Pro, and Ultra will enjoy extended limits, although specific durations have not been disclosed.
Lyria 3 marks the first public release of the model, following its predecessors that were primarily available through Google’s Music AI Sandbox, designed to explore AI’s role in music creation. Previous models were also utilized by select YouTube creators, enabling them to convert speech into music. Google asserts that this latest iteration improves upon its predecessors by addressing several performance issues. Among these advancements, the model now generates its own lyrics, thus simplifying the composing process for users. Additionally, Lyria 3 provides enhanced control over critical song elements, including style, tempo, and vocals, and is capable of producing tracks that are “more realistic and musically complex.”
In a novel twist, users can either input a text-based prompt or upload images and videos for the model to create music based on visual cues. Each generated track will be accompanied by AI-generated album art, created using Google’s Nano Banana model. The company emphasizes that the purpose of these tracks is not to produce musical masterpieces but to offer users an engaging and unique self-expression tool. This approach may suggest that not all generated music will rise to critical acclaim.
In an effort to maintain transparency regarding AI-generated content, Google has embedded all outputs from Lyria with SnythID, a watermarking system. Although this identifier will not be immediately visible, users will have the option to upload audio clips to Gemini, allowing the app to verify the presence of SnythID. This function enables users to detect AI-generated tracks, even if they have not utilized Gemini for creation.
The introduction of Lyria 3 aligns with broader trends in the music industry, where AI technologies are increasingly being utilized to democratize music creation and expand artistic possibilities. As generative models continue to evolve, they have the potential to reshape how music is composed, produced, and consumed. The ongoing integration of AI tools presents both opportunities and challenges, as artists navigate the implications of machine-generated content in a historically human-centric field.
As Google rolls out Lyria 3, it will be interesting to observe not only user engagement with this innovative tool but also how it fits into the larger context of AI’s role in creative industries. The balance between human artistry and machine assistance will likely spark discussions about authenticity, ownership, and the future of music creation.
See also
NVIDIA Researchers Reveal Uniform-State Diffusion Surpasses Masked Models in Reasoning Tasks
Tavus Launches Phoenix-4 Model with Real-Time Emotional Intelligence and Sub-600ms Latency
Top 10 AI Image Generators of 2026: Tools for Stunning Visuals Unveiled
Google Launches Lyria 3 in Gemini App for AI Music Creation from Text and Images
Army Integrates AI in Doctrine Development, Acknowledges Technology’s Critical Flaws
