Gemini Omni Flash

Video Models

generateanimateedittext to videoimage to videovideo to videogooglegeminiomni flashaudio

Overview

Google Gemini Omni Flash video via GCP (Interactions API). One connected image produces image-to-video; two to six images produce reference-to-video; a source video enables edit. Current GCP API exposes 720p output; target duration 3-10s is steered through the prompt for non-edit generations.

Optional first frame (S) and multi-connect Reference Images (R; up to 6 images total), or one source video for edit. Reference them in the prompt as @Image1, @Image2, etc. Text-to-video needs no inputs. Source video max 30s; GCP edit does not accept explicit duration, aspect ratio, or resolution controls.

First frame
Reference Images
Video