Native 720p
1280×720 landscape or 720×1280 portrait. H.264 MP4.
Try Gemini Omni 1.2 Flash online or via API. 3–15 second 720p clips from a prompt or up to 3 images. Billed per second. Failed jobs are free.
| Specification | Credits | Price (USD) |
|---|---|---|
| 720p video | 7per second | $0.07per second |
Prices follow the model configuration. Reference images are not billed unless the model says so.
One request returns one 720p MP4 between 3 and 15 seconds long.
1280×720 landscape or 720×1280 portrait. H.264 MP4.
Pick a whole number of seconds. The default is 8.
Write a prompt alone, or add 1, 2, or 3 stills to guide the look.
What you set on this page is what POST /api/openapi/submit sends.
Generated through the AIReiter OpenAPI on 18 November 2026.
Prompt: A golden retriever running along a sunny beach, cinematic tracking shot, realistic
Request: no images, 16:9, 8 seconds.
Delivered: H.264 MP4, 1280×720, 8.00 seconds.
The live table on this page is the source of truth. Failed jobs are not charged.
The charge follows the length you pick. Longer clips cost more.
Text only, 1, 2, or 3 images: the same length costs the same.
16:9 and 9:16 cost the same.
Submit returns estimated credits for the length you sent. Query returns credits_used when the job completes.
No mode tabs. One prompt, one optional image field.
Leave the image field empty. The prompt drives the whole clip.
Images guide the subject and style. They are references, so the clip does not have to open on your first image.
3–10 seconds: 16:9 or 9:16. 11–15 seconds: 9:16 only. The playground switches to 9:16 for you.
Reference video, video extension, and video editing are not available for this model.
One async contract. The model ID for this page is gemini_omni_1_2_flash.
Send gemini_omni_1_2_flash.
POST /api/openapi/submit with model, params (prompt, aspect_ratio, video_length, image_url), and a unique out_task_id.
POST /api/openapi/query with the same out_task_id. A completed job includes credits_used and the video URL.
A 4th image. 11–15 seconds with 16:9. Lengths outside 3–15. A reference video (video_url).
Need audio, 1080p, or a reference clip? These run in the same playground.
Text, multi-image reference, first/last frames, and reference video up to 15s.
02Wan 3.0Text, first/last frames, mixed references, and document-to-video up to 30 seconds.
03Veo 3.1Google's video model with native audio. Fast and Lite variants for quicker runs.
04HappyHorse 1.1Reference-to-video from 1–9 ordered images, 3–15 seconds, up to 1080p.
| Control | Type | Default | Notes | |
|---|---|---|---|---|
| 01 | prompt | text | — | — |
| 02 | image_url | image | — | Optional. Up to 3 images. Leave empty for text-to-video. Reference video is not supported. |
| 03 | aspect_ratio | string | 16:9 | 16:9 landscape or 9:16 portrait. 11–15 seconds is portrait only. |
| 04 | video_length | string | 8 | 3–10s landscape or portrait. 11–15s portrait only. |
Subject first, then camera and light. Skip images for text-to-video.
Product shots, a character, or a style frame. Three is the cap.
3 to 15 seconds. 16:9 or 9:16 up to 10 seconds; 9:16 from 11 seconds.
POST /api/openapi/submit returns a quote. Poll POST /api/openapi/query with the same out_task_id.
gemini_omni_1_2_flash
POST to /api/openapi/submit with the model, a unique out_task_id, and parameters.
POST to /api/openapi/query with the same out_task_id until the task completes.
Save outputs in your own storage rather than hot-linking them.
No. Every clip is native 720p: 1280×720 landscape or 720×1280 portrait.
Billed per second of output, from 3 to 15 seconds. Images and aspect ratio do not change the price. Failed jobs are not charged. The table on this page is current.
No. This model takes a prompt and up to 3 images. For reference video, use Kling v3 Omni.
Any whole number from 3 to 15 seconds. The default is 8.
The model only renders 11–15 second clips in 9:16. The playground switches to 9:16 for you; the API rejects 16:9 at those lengths.
No. Your images are references for subject and style. The clip is not forced to start or end on them.
gemini_omni_1_2_flash. Send it to POST /api/openapi/submit and poll POST /api/openapi/query with the same out_task_id.
No. This page runs Gemini Omni 1.2 Flash through the AIReiter API with AIReiter credits. Google's own Gemini docs are on ai.google.dev.