Google says Gemini Omni 1.2 Flash makes video up to 15 seconds, accepts image, short motion-clip and audio references, and supports multi-turn editing. The prompts below are our own and untested at scale, so treat them as starting points. Google also names fast sports action, rapid script changes and complex physics as areas still being improved, so avoid leaning on those.
Update, November 10, 2026: Google says Omni 1.2 Flash is also in Google Vids at no cost, up to three 15-second videos a month. See the summary.
Plan the 15 seconds
- Write one scene with a beginning, middle and end, for example reveal, action, close-up.
- Name the camera move once. Google says camera control is better and jump cuts are fewer, so ask for a single continuous shot.
- If someone speaks, write the line in quotes and say who says it.
Single-take product clip
One continuous 15-second shot. A matte ceramic coffee mug on a wooden table by a sunny window. The camera starts wide, slowly pushes in as steam rises, and ends on a close-up of the mug's handle. Soft morning light, quiet cafe ambience, no cuts, no text.Reference-guided scene
Upload a character or product image first, then:
Use the uploaded image as the main character. Keep the outfit and hair exactly the same. She walks through a night market, stops at a noodle stall and says "This one smells amazing." Warm lantern light, handheld camera, 12 seconds.Multi-turn edit
After the first clip, change one thing at a time:
Keep the same scene and camera move. Change the time of day to golden hour and slow the pacing slightly.Draft first, then upscale
Google suggests drafting at 360p or 720p and upscaling to 1080p or 4K. Test the idea cheaply, adjust the prompt, then upscale only the clip you like.
Checklist
- One scene, one camera move.
- Quote any spoken lines.
- Avoid fast sports and complex physics.
- Check Google’s pricing before generating at scale.
More on structure in how to write AI video prompts. Compare with Veo in Veo vs Kling.