Video Generation

Google Veo models power text-to-video through /v1/video/generations. For the full walkthrough covering every provider, see Video Generation.

Example

curl https://llmtr.com/v1/video/generations \
  -H "Authorization: Bearer llmtr-your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/veo-3.1-generate-001",
    "prompt": "A cinematic drone shot over Cappadocia at sunrise",
    "duration_seconds": 8,
    "aspect_ratio": "16:9",
    "resolution": "1080p",
    "max_wait_seconds": 180,
    "max_budget_usd": 3.00
  }'

Parameters

Field	Description
`model`	Canonical id of the video generation model. Required.
`prompt`	Text description of the video to generate. Required.
`duration_seconds`	Video duration. Required. Veo supports 4-8 seconds.
`aspect_ratio`	Aspect ratio. Defaults to `16:9`.
`resolution`	`720p`, `1080p`, or `4k`. Veo 3.1 Lite does not support 4k. Defaults to `720p`.
`negative_prompt`	Description of elements to avoid in the video.
`image_url`	Image-to-video reference frame. Only the `data:image/...;base64,...` form is accepted; a remote URL returns 400 `unsupported_input`.
`max_wait_seconds`	Polling timeout for the long-running job. At most 300 seconds.
`max_budget_usd`	Spend cap the gateway checks before starting the request.

Response

Veo returns the video as base64 MP4 in data[].b64_json; the url field is null for these models.

{
  "id": "projects/.../operations/...",
  "object": "video.generation",
  "model": "google/veo-3.1-generate-001",
  "status": "completed",
  "data": [
    {
      "url": null,
      "b64_json": "AAAAIGZ0eXBpc29t...",
      "mime_type": "video/mp4"
    }
  ]
}

Pricing

Veo models are billed per second by native unit; the per-resolution price table lives on the Video Generation page. Always cap with max_budget_usd.

Discontinued endpoints

Google shuts down the pre-3.1 Veo endpoints on 28 August 2026. The successor for google/veo-3.0-generate-001 and google/veo-2.0-generate-001 is google/veo-3.1-generate-001; for google/veo-3.0-fast-generate-001 it is google/veo-3.1-fast-generate-001. After the cutoff, requests for the old identifiers return 410 model_retired; there is no auto-forward.