Video and sound together
Every Omni Flash clip comes with an audio track: the voice, the room and the small sounds of the scene, generated with the picture rather than added after.
Gemini Omni Flash AdAnt · by Google
Gemini Omni Flash is Google's fast video model with native audio, and you can try it here: type a prompt and get a short clip with sound in under a minute.
We ran this prompt once on each model on AdAnt, with no retries and no edits: 9:16, 720p, 6 seconds, audio on. Whether she says the line clearly and in sync, whether the bottle keeps its shape in her hand, and whether it reads as a phone video rather than a film.
Handheld selfie-style UGC video. A woman in her late 20s with curly brown hair, in a sunlit bathroom, holds a small amber glass serum bottle up next to her face, smiles at the camera and says: "This is the only serum I've ever finished twice." Natural window light, real skin texture, slight handheld shake, vertical 9:16 phone framing.
| Model | Made by | Modes | Longest clip | Resolution | Audio | Inputs | Test run | Test cost | Best for |
|---|---|---|---|---|---|---|---|---|---|
| Gemini Omni Flash | Text to video, image to video | 10 s | 4K | Native | Start/end image, up to 30 reference images | 37 s | about 3 credits | fast drafts and talking clips with sound | |
| Seedance 2.5 | ByteDance Seed | Text to video, image to video, reference to video | 30 s | 4K | Native | Start/end image, up to 30 images, 10 videos, 10 audio clips | 2 min 36 s | about 9 credits | longer story clips and complex scenes with many references |
| Seedance 2.0 | ByteDance Seed | Text to video, image to video, video editing | 15 s | 4K | Native | Start/end image, up to 9 images, 3 videos, 3 audio clips | 4 min 7 s | about 6 credits | UGC ads and product clips with references |
| Seedance 2.0 Fast | ByteDance Seed | Text to video, image to video | 15 s | 4K | Native | Start/end image, up to 9 images, 3 videos, 3 audio clips | 1 min 43 s | about 5 credits | testing many hooks and ideas cheaply |
| Topview T1 | Topview | Text to video, image to video | 30 s | 1080p | Native | Start image, optional end image | 3 min 19 s | about 5 credits | longer single takes and calm, natural clips |
Test run and cost come from the same-prompt run above, measured on AdAnt. Longer clips and higher resolutions cost more; the try-it panel shows the price before you run.
Every Omni Flash clip comes with an audio track: the voice, the room and the small sounds of the scene, generated with the picture rather than added after.
Our 6-second test clip was ready in 37 seconds, about three times faster than Seedance 2.0 Fast and seven times faster than Seedance 2.0 Pro.
The test clip cost about 3 credits, the lowest of the video models on AdAnt, so trying five versions of a hook costs less than one long clip elsewhere.
Lock the opening and closing image, or pass up to 30 reference images for the people, products and places in the scene.
Handheld selfie video. [Who, in a few words] in [where], holds up [your product] and says: "[one short line]". [Light], [one camera detail], vertical phone framing.
Gemini Omni Flash is Google's fast multimodal video model. It turns a prompt, and optionally start and end images or reference images, into a short video with its own audio. On AdAnt it runs as Gemini Omni 1.1 Flash for clips of 3 to 10 seconds, from 360p drafts up to 4K.
Make several openings of the same ad quickly and cheaply, then keep the one that stops the scroll.
A person holding a product and saying one line, with the sound already in.
Turn a shot list into rough clips before you spend on a final model.
Gemini Omni Flash is Google's fast multimodal video model. It turns a text prompt, and optionally start and end images or up to 30 reference images, into a short video with its own audio track, speech included. In our same-prompt test on AdAnt it made a 6-second clip in 37 seconds, the fastest of the video models we ran.
Running this model uses paid credits from an AdAnt plan. Creating an account is free. A 5-second 720p clip costs about 3 credits.
Omni Flash for speed and cost: our test clip took 37 seconds and about 3 credits. Seedance 2.5 for longer clips, up to 30 seconds, and the most references: the same test took 2.5 minutes and about 9 credits. Watch both on this page.
Yes. Every clip has an audio track generated with the video, including speech.
From 3 to 10 seconds per clip on AdAnt. AdAnt's agents stitch several clips for a longer ad.
Google offers its Gemini models to developers through the Gemini API. On AdAnt you can run it without your own provider key, from the adant-cli media command or the AdAnt plugin for Claude, paying in AdAnt credits. Our 6-second test clip cost about 3 credits.
VideoByteDance Seed · longer story clips and complex scenes with many references
VideoByteDance Seed · UGC ads and product clips with references
VideoByteDance Seed · testing many hooks and ideas cheaply
VideoTopview · longer single takes and calm, natural clips
MotionKuaishou (Kling AI) · putting a consistent character into a proven performance
ImageGoogle DeepMind · photoreal product shots and keeping a product on brand
ImageGoogle DeepMind · fast social images, thumbnails and drafts
ImageOpenAI · posters, infographics and anything with words
AdAnt's agents research your market, write the script, pick the right model for each shot and make the video in one chat.
Start with AdAnt