Google DeepMind launched Lyria 3 on February 18, 2026. The tool is integrated into the Gemini application and can now be used by people aged 18 and over around the world. Ela generates high-fidelity audio tracks lasting up to 30 seconds from simple commands.
The technology accepts text, image or video prompts as a basis for compositions. Results include vocals, lyrics, and automatically generated cover art. The launch took place on the same day that Brazilian athletes broke South American records and tested the tool in practice.
The demonstration involved news of the records in the 60 meter dash in São Paulo. The AI’s response was a vibrant trail that paid homage to the sprinters’ achievement.
Immediate access to the tool in the Gemini app
Gemini users aged 18 or over can find the resource directly in the app’s interface. Activation takes place without the need for extra installation and is initially available in the desktop version. The expansion to mobile devices takes place in the days following launch.
The company confirms that access follows global privacy and responsible use rules. Qualquer person can try track generation at no additional cost at the time of beta launch.
Introducing Lyria 3, our new music generation model in Gemini that lets you turn any idea, photo, or video into a high-fidelity track with custom lyrics.
—Google Gemini (@GeminiApp)February 18, 2026
From funny jingles to lo-fi beats, you can create custom 30-second soundtracks for any moment.
See how it works. 🧵
Technical features of the model Lyria 3
The Lyria 3 processes text commands to define the style, rhythm and mood of the track. The system interprets detailed descriptions and produces audio with a natural flow between notes.
Images and videos serve as a visual reference for musical creation. The model analyzes visual elements and translates the content into corresponding sounds.
The tool supports multiple languages in the generated lyrics. Exemplos include Portuguese, English and Hindi with natural pronunciation.
The end result includes an invisible watermark for identifying AI-generated content. The company applies SynthID to all tracks produced.
Practical demonstration with Brazilian athletics records
A user inserted the news about the South American records in the 60 meter dash. The tool responded within a few seconds with the track called “Feet Velozes”.
Ana Azevedo completed the women’s race in 7.05 seconds during Circuito Performance from Atletismo Short Track to São Paulo. The brand surpassed the previous record and secured an index for Mundial Indoor.
Erik
The AI created an energetic composition that captured the spirit of speed and conquest. The finished audio could be played immediately within the application.
How Lyria 3 processes text and media prompts
The model analyzes user-supplied text for keywords related to emotion and rhythm. It then combines these elements with pre-trained musical patterns.
When receiving images or videos, the system extracts visual characteristics such as colors, movement and environment. Esses data guides the choice of instruments and track tempo.
The user can refine the result with additional commands such as “faster” or “add female vocals”. The tool responds to changes in real time.
Complete generation takes just a few seconds even on complex prompts. The final audio has quality comparable to short-term professional productions.
Initial reactions from users and composers to the news
Several users reported ease of using the Lyria 3 shortly after launch. Muitos highlighted the speed in generating personalized tracks for personal or professional projects.
Professional composers have expressed doubts about the model’s training sources. DeepMind said it only uses material licensed from YouTube and commercial partners.
The company reinforces that the training respects copyright and current legislation. Representantes state that the tool aims to assist and not replace human creators.
The debate about ethics in musical AI continues to be active in the artistic community. Especialistas monitor developments to assess future impacts on the sector.
Availability and requirements for use of Lyria 3
The tool requires an active account on Google and a minimum age of 18 to access. The resource appears directly in the Gemini interface without additional configuration.
Initially available in the web version of the application, mobile support arrives in a few days. Usuários from any country can test as long as they meet the age criteria.
DeepMind plans regular updates to improve quality and add new musical styles. Feedback from early adopters guides future improvements.
Integration with other Google services like YouTube Shorts makes it easy to share created tracks. The export process is simple and straightforward.
The company maintains constant monitoring to ensure compliance with data privacy and security standards. Qualquer necessary adjustment is implemented quickly.
Creative applications of technology in different areas
Lyria 3 opens up possibilities for content producers who need original tracks in the short term. Criadores of social networks use the tool to sound videos and stories.
Educators explore the resource to demonstrate musical concepts in an interactive way. Estudantes generate practical examples during composition and production classes.
Independent artists test ideas quickly without the need for a full studio. AI serves as a starting point for further manual refinement.
Advertising companies create personalized jingles from visual briefs. The process speeds up the production of digital campaigns.
Integration with the Google ecosystem
The Lyria 3 naturally connects to the Google ecosystem to facilitate workflow. Usuários transfer tracks directly to video publishers or streaming platforms.
Cover art generated by Nano Banana visually complements each track produced. The end result forms a complete package ready for publication.
DeepMind plans to expand the model’s capabilities based on users’ actual usage. Atualizações future tracks may include longer tracks and more precise controls.
The launch marks another step in the evolution of creative tools accessible to everyone. Technology democratizes high-quality music production.
The community looks forward to the next innovations that will emerge from this solid foundation. The focus remains on the user experience and the quality of the audio delivered.
