Keep pulling the thread on Khulan Davaajav.
Google's Gemini model can analyze a video's content frame-by-frame to generate a detailed, timestamped prompt for the Lyria 3 Pro model to create a synchronized soundtrack.
Google launched Gemini 3 Flash Live with a Live Avatar feature in preview on the day of the Google Cloud Next '26 presentation.
The Gemini 3 Flash Live model with the Live Avatar feature is connected to Google Search, enabling it to provide answers based on live data.
Google has a world model named Journey 3, which is not yet available on Google Cloud.
Google's Generative Media Models is an umbrella term for its portfolio of creative AI models, including Nano Banana, Vio, Gemini Audio, and Lyria.
Google is releasing new updates for its generative media models approximately every one to two weeks.
Google's Nano Banana model allows for detailed artistic control in prompts, including specifying camera types, lens types, glossy highlights, halation, and luminance.
The Gemini app can analyze images to help users generate detailed prompts describing artistic style, texture, and camera settings for use in other generative models.
Google's Vio 3.1 Lite is its most cost-effective video generation model currently on the market.
Google's Vio 3.1 Lite model can generate most video frames in under 60 seconds.
Google's Vio 3.1 Lite model supports a "first frame, last frame" feature to control the beginning and end points of a generated video clip.
Google's Lyria 3 Pro model can understand and incorporate specific timestamps within a prompt to control the musical composition over time.