Gemini Omni 1.2 is Google's multimodal video model, live on reAPI. Gemini Omni 1.2 makes 3–15 second clips with native sound from a prompt, an image, first and last frames or subject references, then edits or extends a clip with a plain-language instruction, at 360p, 720p, 1080p or 4K.
Required in every mode
360p / 720p / 1080p / 4k · default 720p
3–15 s · default 10 · at most 10 with two frames or for an extension
16:9 / 9:16 · default 16:9
Try one of these prompts
Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.
One 5-second sample at the cheapest tier.
Gemini Omni 1.2 is the newest version of Gemini Omni Flash, Google's video model. Google describes it as a multimodal model built for fast video generation, conversational editing and cinematic control: Gemini Omni 1.2 reads text, images and video, and returns a clip with synchronized native audio of up to 15 seconds. On reAPI, Gemini Omni 1.2 runs under the model id gemini-omni-1-2 on the same video endpoint as every other model. One Gemini Omni 1.2 request can start a clip from a prompt, an image or two frames, keep characters on-model from reference images, or take an existing clip and change one thing in it. This Gemini Omni 1.2 clip was made from a first and a last frame.
Mode is implicit: which media fields you send decides how the request runs.
promptA prompt alone. Gemini Omni 1.2 writes the shot, the motion and the sound.
first_frame_image + last_frame_imageStart from one image, or give a first and a last frame and Gemini Omni 1.2 animates the move between them.
image_urls + image_with_rolesIn Gemini Omni 1.2, several images of a character, an outfit or a product keep them consistent across the shot.
video_urls / extend_from_task_idSend Gemini Omni 1.2 a clip of up to 15 seconds, or name an earlier task, and change one thing in it or add up to 10 more seconds. A Gemini Omni 1.2 edit bills the source length, an extension the added seconds, each plus an eighth of the source.
Real-world workflows and production use cases you can build and ship with this model.
Describe the subject, the camera and the sound, and Gemini Omni 1.2 returns a finished clip with audio. Google's guide notes Gemini Omni 1.2 likes to cut between shots, so ask for one continuous shot when you need it. This marble run is one 8-second Gemini Omni 1.2 generation at 720p from the prompt in the playground above.
Give Gemini Omni 1.2 a first frame and a last frame, and Gemini Omni 1.2 animates the transition between them in 3 to 10 seconds. That fixes both ends of a Gemini Omni 1.2 shot, which a single start image cannot. This 5-second Gemini Omni 1.2 clip moves a lighthouse keeper from the window to a brass lantern between two still frames.
Send a clip of up to 15 seconds with a short instruction, such as "turn the rain into snow, keep everything else the same", and Gemini Omni 1.2 changes that and preserves the rest of the clip. A Gemini Omni 1.2 edit keeps the source length. This 3-second rainy street with a paper boat came back from Gemini Omni 1.2 covered in snow.
Gemini Omni 1.2 makes a clip, edits it, and continues it, so a shot can go from first draft to final cut without moving to another model. Extending an earlier Gemini Omni 1.2 task needs only its task ID.
Your reAPI key already covers Gemini Omni 1.2, Veo, Seedance, Kling, Wan and the rest. It is the model id gemini-omni-1-2 on the same /api/v1/videos/generations endpoint, with no Google Cloud project or separate billing account.
reAPI bills Gemini Omni 1.2 by the second at the rate of the resolution you pick, an edit or extension also bills an eighth of its source clip, credits never expire, and failed generations are refunded automatically. Current rates are in the pricing table on this page.
Gemini Omni 1.2 and 1.1 are both live on reAPI. Values are what each model accepts on reAPI. Google routes its own 1.1 requests to 1.2 from December 3, 2026.
Rates for both models are in each page's pricing table. Google's own list price for Gemini Omni 1.2 is lower per second than for 1.1.
Sign up and create a key in the dashboard. New accounts get a small test credit; add credits before your first Gemini Omni 1.2 clip, which costs more than that.
OpenPick a Gemini Omni 1.2 mode, write a prompt, and generate. Copy the request from the API tab when you like the result.
OpenPOST to /api/v1/videos/generations with model gemini-omni-1-2, then poll the Gemini Omni 1.2 task until it completes and download the video.
Opencurl -X POST https://reapi.ai/api/v1/videos/generations \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-omni-1-2",
"prompt": "A glass marble rolls fast along a wooden chain-reaction track through a sunlit workshop, knocking dominoes that spin a tiny windmill, one continuous smooth tracking shot, clacks and soft workshop ambience. No text, no logos.",
"resolution": "720p",
"duration": 8,
"aspect_ratio": "16:9"
}'Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.
401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.
Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.
A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.
Common questions about this model.
Explore more models in the same category.
From $0.347 per generation
From $0.495 per generation
From $0.081 per generation
Kuaishou
From $0.063 per second
Try it in the playground or grab an API key to integrate now.