Google has unveiled Lyria 3, the latest iteration of its AI music generation tool, capable of creating a 30-second audio track in seconds from a variety of inputs—including text prompts, images, and even short video clips. The new model produces full compositions featuring instrumentals, lyrics, and vocal performances, all generated automatically with high audio fidelity. While the AI music space has been dominated by platforms like Suno and Udio since their rapid rise in 2024, Lyria 3 brings a fresh approach by integrating multimodal input. Users can describe a mood, genre, or scene in words, upload a photo that evokes a particular atmosphere, or provide a video clip to inspire a musical piece. The system then synthesizes a complete, coherent track tailored to the input. One key limitation compared to some competitors is the 30-second output length per generation. However, Google emphasizes that the quality of the generated music—particularly the natural-sounding vocals and well-structured arrangements—represents a significant leap forward. The model leverages advanced neural audio synthesis and language understanding to ensure lyrics and melodies align cohesively with the input. Lyria 3 builds on the foundation of earlier versions, improving on vocal clarity, emotional expressiveness, and the overall realism of instrumentals. It also features enhanced control over style and tone, allowing users to fine-tune results based on genre, tempo, and vocal characteristics. Though not yet available to the public, Lyria 3 is part of Google’s broader push into creative AI tools, aiming to empower musicians, content creators, and developers with accessible, high-quality music generation. The model reflects a growing trend in AI: turning diverse forms of media into rich, dynamic audio experiences in real time.