Gemini Omni 1.2 Flash

Use gemini-omni-1.2-flash to create video output. This page contains the complete request needed for this model.

Model facts

Property	Value
Model ID	`gemini-omni-1.2-flash`
Endpoint	`POST /api/v1/generation/create`
Output	Video
Modes	text-to-video, image-to-video, video reference
Pricing	View current pricing

Minimal request


curl -X POST https://api.api-stock.com/api/v1/generation/create \
  -H "Authorization: Bearer $API_STOCK_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light"
  }
}'

The create response contains a task ID:


{
  "code": 200,
  "data": {
    "taskId": "019ca881-9503-7270-a560-f00fc2b15785",
    "status": "not_started",
    "createdAt": "2026-08-22T12:00:00.000Z"
  }
}

Parameters

Parameters belong inside the top-level input object. Unknown fields are rejected.

Parameter	Type	Required	Default	Description
`prompt`	string	Yes	—	Describe the scene, motion, sound, or video edits. max 4000 chars
`imageUrls`	string[]	No	—	First frame and optional last frame. Two frames limit duration to 10 seconds. Cannot be combined with referenceImageUrls or videoUrls.
`referenceImageUrls`	string[]	No	—	Up to 10 subject reference images; prefer at most 6 for character consistency. Cannot be combined with imageUrls.
`videoUrls`	string[]	No	—	One source video, at most 15 seconds, to edit or extend. Editing preserves its length; duration is ignored. Source duration is checked before billing.
`videoMode`	“edit” \| “extend”	No	—	Source-video operation. Extension adds 3–10 seconds and bills only the added segment. Requires videoUrls.
`duration`	integer	No	`10`	Whole output seconds (3–15), or added seconds for extension (3–10). Ignored when editing a source video. min 3, max 15
`resolution`	“360p” \| “720p” \| “1080p” \| “4k”	No	`"720p"`	Output resolution; 1080p and 4K are upsampled.
`aspectRatio`	“16:9” \| “9:16”	No	`"16:9"`	Output aspect ratio

Modes

Minimal generation


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light"
  }
}

15-second video


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "duration": 15,
    "resolution": "720p"
  }
}

First and last frames


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "imageUrls": [
      "https://example.com/start.jpg",
      "https://example.com/end.jpg"
    ],
    "duration": 8
  }
}

Edit a source video


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "videoUrls": [
      "https://example.com/source.mp4"
    ],
    "videoMode": "edit"
  }
}

Extend a source video


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "videoUrls": [
      "https://example.com/source.mp4"
    ],
    "videoMode": "extend",
    "duration": 4
  }
}

With image urls


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "imageUrls": [
      "https://example.com/input.jpg"
    ]
  }
}

With reference image urls


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "referenceImageUrls": [
      "https://example.com/input.jpg"
    ]
  }
}

With video urls


{
  "model": "gemini-omni-1.2-flash",
  "input": {
    "prompt": "A cinematic scene with soft natural light",
    "videoUrls": [
      "https://example.com/input.mp4"
    ]
  }
}

Constraints

prompt: maximum 4000 characters.
videoMode: allowed: edit, extend.
duration: minimum 3; maximum 15.
resolution: allowed: 360p, 720p, 1080p, 4k.
aspectRatio: allowed: 16:9, 9:16.
Text, a single first frame and subject references support 3–15 whole seconds (default 10). First/last-frame interpolation and extension support only 3–10 seconds.
imageUrls holds the first and optional last frame; it cannot be combined with referenceImageUrls or videoUrls. Up to 10 subject reference images or one source video of at most 15 seconds are accepted.
With videoUrls, editing is the default: duration is ignored, the source length is probed before billing and rounded up to whole seconds. Extension requires videoMode extend and bills only the added seconds; the returned video also includes the original source.
Audio is generated from the prompt. 1080p and 4K are upsampled. Previous provider task IDs and conversational reuse are not accepted; supply a public source-video URL.

Invalid types, unsupported enum values, out-of-range numbers, and missing required fields return HTTP 400 with error.code: "invalid_input". Correct the request rather than retrying it unchanged.

Result

Poll the task or provide a top-level webhook URL in the create request:


curl https://api.api-stock.com/api/v1/task/status/$TASK_ID \
  -H "Authorization: Bearer $API_STOCK_KEY"

A successful media task returns one or more files:


{
  "code": 200,
  "data": {
    "taskId": "019ca881-9503-7270-a560-f00fc2b15785",
    "status": "finished",
    "files": [
      {
        "fileUrl": "https://s3.api-stock.com/result.mp4",
        "fileType": "video"
      }
    ],
    "createdTime": "2026-08-22T12:00:00.000Z"
  }
}

See Tasks and webhooks for terminal statuses, retries and webhook delivery.