Google has unveiled its AI music generation model, Lyria 3 Pro, TechCrunch reported on Tuesday. The new version extends the maximum length of generated tracks to 3 minutes from 30 seconds and lets creators specify elements such as intros, verses and choruses in more detail. It will be offered first to paying subscribers on the Gemini app and will also be applied to Google’s video editing app and ProducerAI, an AI music production tool Google recently acquired.

Google has unveiled its AI image-generation model \'Nano Banana 2\' (Gemini 3.1 Flash Image), TechCrunch reported on Feb. 26. The model can generate images in various aspect ratios from 512px up to 4K and can express up to five characters and 14 objects. It will be used as the default model across Gemini image modes and Flow, and as Google Search’s default image generator in 141 countries. API support is also expanding.