NezhaGateNezhaGate
Video ● Operational Service status →

Gemini Omni Flash

⧉
gemini-omni-flash

Gemini Omni Flash is Google's next-generation video model, released in 2026. One endpoint, parameters pick the mode: text-to-video with no image, one image as the first frame, two as first and last frame, and with image_role=reference one to three images as subject references. 6, 8 or 10 seconds, 720P or 1080P, landscape 16:9 or portrait 9:16, with native audio, and strong at following a storyboard across several shots. Asynchronous: submit, then poll the job id for a re-hosted mp4 link (kept for 60 days). Billed per second at $0.05/s; failed jobs are refunded in full.

Text-to-videoImage-to-videoFirst/last frame1080P

Sample clips

All made with NezhaGate. Click a clip to hear it.
Beneath the Walls of Troy HD original
Myth Awakens HD original
Grandma's Birthday HD original

Live Test · Playground

Try out Gemini Omni Flash right here (available after login).

Input

0 / 20000 chars
Allowed lengths: 720P 6/8/10s · 1080P 6/8/10s
Reference images0 / 3
Click to upload or drag files here jpg / png / webp / gif, up to 16 MB each or paste links
With multiple references, address them in the prompt as @Image1 / @Image2, e.g. "the character from @Image1 stands in the scene from @Image2". Without tokens the model decides on its own.
Uploaded references are kept for 7 days.
Rendering usually takes 5-20 min; when the backend is busy it can exceed 60 min (we wait up to 90). Poll the job id rather than resubmitting - each resubmission is billed separately. Billed per second, fully refunded on failure. Prompt limit 20000 characters (Chinese and English count the same).

Output

video
Results will appear here after running.
🕑 Results are kept for 60 days and then deleted automatically, so download anything you want to keep.

About Gemini Omni Flash

Gemini Omni Flash is Google's next-generation video model, released in 2026. On NezhaGate one model and one set of parameters cover text-to-video, first-frame video, first-to-last-frame transitions and multi-image references: no image is text-to-video, one image is the first frame, two are the first and last frame, and with image_role=reference one to three images act as subject references. 6, 8 or 10 seconds, 720P or 1080P, landscape 16:9 or portrait 9:16, with native audio. It is strong at cutting between the shots a prompt describes, so a 10-second clip can tell a whole short beat. Asynchronous, billed per second at $0.05/s, failed jobs refunded in full.

Use cases

Storyboarded shorts

Write the framing, action and sound shot by shot in the prompt; the model cuts between them in order and tells a short beat within 10 seconds.

Characters from references

Give a character, prop or set with one to three reference images and place them in new scenes, ideal for series content.

First-to-last-frame transitions

Provide a starting and an ending image and get the camera move and transition in between, for example an aerial shot that descends into a street.

Vertical social content

9:16 portrait at 1080P, ready for short-video platforms; 6 seconds suits fast-paced cuts, 10 seconds a complete mini story.

How to choose

Telling a short multi-shot beat, or want the model to follow the storyboard in your prompt closely? Choose Gemini Omni Flash; for a single 8-second shot Veo 3.1 works too. The price follows the length (6 s $0.30, 8 s $0.40, 10 s $0.50) and 720P and 1080P cost the same, so pick 1080P for final delivery. For clips longer than 15 seconds use Seedance 2.5 or Wan 3.0.

FAQ

How is Gemini Omni Flash different from Veo 3.1?
Both come from Google. Omni Flash is the 2026 next-generation model, better at cutting between storyboarded shots and multi-shot storytelling, with 6, 8 or 10 seconds billed per second; Veo 3.1 is always 8 seconds and billed per clip.
How do I use reference images?
One image is the first frame by default and two are the first and last frame. To make the people or objects in the images appear in the video rather than serve as a frame, pass image_role=reference with up to three images.
Which lengths, resolutions and aspect ratios are supported?
6, 8 or 10 seconds, 720P or 1080P (same price), 16:9 or 9:16. Any other value returns an error instead of being changed silently.
How is it billed? Am I charged for failures?
Per second at $0.05/s: the chosen length is held at submit and the same amount is settled on success. A failed render, a content-policy block or a timeout is refunded in full.

Related Models

Explore other models you can integrate.

View all →