Gemini Omni Video is a multimodal AI video generation model from Google. Combine text prompts with reference images (up to 7), narration voices (30 options), and an optional source video clip (10-second slice) to produce 4 / 6 / 8 / 10-second outputs in 16:9 or 9:16. Ideal for storyboard-driven generation where you want fine control over scene, character, and narration in one pass.