On February 18, 2026, Google announced that the Gemini app had begun rolling out Lyria 3, Google DeepMind’s newest music-generation model. In beta, it lets users create 30-second music tracks by typing a prompt or uploading an image, so a short text idea—or even a single photo—can quickly become a custom soundtrack. Google says the system can also generate cover art for the track, turning the experience into a small but complete creative package. (blog.google)
What makes Lyria 3 especially interesting is that it does more than simply produce background noise. Google says the model can generate lyrics automatically, give users more control over style, vocals, and tempo, and create more realistic, musically complex results than earlier Lyria versions. DeepMind also describes Lyria 3 as a high-fidelity music tool built with input from producers and musicians, and its prompt guide shows that users can start with a simple idea like “a rock song” or become more detailed by specifying instruments, dynamics, and vocal style. In other words, the tool rewards imagination, but it also rewards precise language. (blog.google)
For Japanese learners of English, that last point is part of the fun. If you ask for “a calm piano ballad,” you may get one result; if you ask for “a dreamy city-pop track with soft female vocals and a summer-night mood,” you may get something much more vivid. Because image prompts are supported too, a sunset photo, a café scene, or a travel snapshot can become the starting point for music. This means English is not only something to study here—it becomes a creative tool for shaping sound, mood, and storytelling. (blog.google)
There is also a clear safety message behind the feature. Google says all tracks generated in the Gemini app are embedded with SynthID, its watermarking technology for identifying Google AI-made content, and Gemini’s verification tools now extend to audio as well as images and video. As of the rollout, Lyria 3 music generation was available for users aged 18 and over in eight languages, including Japanese, with higher usage limits for Google AI Plus, Pro, and Ultra subscribers. AI music is becoming easier than ever—but with Lyria 3, Google is also trying to make it more transparent and more responsible. (blog.google)
