Google Launches Music Generation Model to Make Songs 30-Seconds at a Time
DeepMind's latest model generates lyrics and music from text or images, but free users are capped at 30 seconds.
Deep Dive
Google's DeepMind lab has launched Lyria 3, its latest music generation model, within the Gemini app. The beta feature creates 30-second tracks with AI-generated lyrics and album art from text or visual prompts, offering control over style and tempo. Free users are limited to 30-second clips, while paid subscribers get higher limits. All outputs are watermarked with SynthID for AI detection.
Why It Matters
Democratizes music creation for casual users and integrates AI-generated audio into the mainstream creative workflow.