Gemini Omni Flash is Google's preview video-generation model that renders short, native-audio video clips from a text prompt or a reference image. In UGC Copilot it is the fifth video engine (alongside Sora 2, Veo 3.1, Kling 3.0, and Seedance 2.0), offered as a labeled preview for fast drafts.
What makes Omni Flash different
Unlike the other engines, Omni Flash is built on Google's Interactions API rather than a one-shot generation call. It processes text, image, and audio together, so a single request produces video with native audio — voiceover or synced dialogue described in the prompt — in one step. When you provide a scene image, Omni uses it as the video's start frame, giving exact likeness (the on-camera person matches your generated creator), the same way Veo 3.1 and Kling 3.0 do.
Capabilities and limits
Omni Flash renders 720p only (there is no HQ tier), in 4–10 second clips, in 16:9 or 9:16 aspect ratios. Renders are fast — typically under a minute for an 8-second clip. Because it is a preview engine, its resolution, clip length, and pricing can change, which is why it is positioned for drafts and iteration rather than final, client-facing ads.
When to use it
Reach for Omni Flash when you want a quick, Google-native draft with audio baked in and 720p is acceptable. For polished 1080p output or a dedicated HQ tier, Veo 3.1 is the stronger pick; for the lowest cost per second on faceless product ads, Sora 2 wins. Most UGC Copilot users draft with one engine and ship on another — Omni Flash is a strong drafting option.