Grok Imagine video API: image to video, 4–15 seconds

Yes, you can call Grok Imagine video over an API. On Sume's video API the model is grok-imagine-video-1.5 (Grok Imagine Video 1.5) on POST /v1/videos, and it runs as image-to-video only: you send a first-frame image and a prompt, and it returns a 4–15 second clip at 480p or 720p, with no audio track. On Sume it does not generate from text alone.

The model limits come from the grok-imagine-video-1.5 catalog entry behind GET /v1/videos/models and the Sume API reference; the request flow comes from the Video generation docs. All were read on 2026-09-28, and "current code" marks behavior read from the API code.

How do I call the Grok Imagine video API?

Put the image in frame_images as the first_frame and describe the motion in prompt. The image must be a public HTTPS URL. The legacy Video Router route takes the same image as image_url, which its reference marks "Required for grok-imagine-video-1.5".

curl -X POST "https://api.sume.com/v1/videos" \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: grok-pushin-001" \
  -d '{
    "model": "grok-imagine-video-1.5",
    "prompt": "Subtle camera push-in with natural motion.",
    "frame_images": [
      { "type": "image_url", "image_url": { "url": "https://example.com/first-frame.png" }, "frame_type": "first_frame" }
    ],
    "resolution": "720p",
    "duration": 6
  }'

What does grok-imagine-video-1.5 accept?

Fewer fields than most catalog models. In current code, a value the model doesn't list for these fields is refused at submit with 400 unsupported_capability.

Can Grok Imagine make a video from text only?

Not on Sume. The catalog marks the model text_to_video: false, and its constraints read "image_url or first_frame_url required". For a clip from words alone, pick a model that lists text-to-video; Text to video API explains that request, and Kling 3 vs Grok Imagine sets this model beside kling-3.

What shape will the clip be?

The model takes no aspect_ratio, so crop the first frame to the shape you want, such as 9:16 for a vertical clip, before you send it. Vertical 9:16 video generation API lists which models take the field.

How much does the Grok Imagine video API cost?

Sume bills video at the provider's list price × 1.25, reserved when you submit. grok-imagine-video-1.5 has one per-second rate, $0.0125, at both 480p and 720p, so a 10-second clip reserves $0.13 and a 15-second clip $0.19, plus a 5.5% agent fee by default. pricing_skus shows it as per-video-second, and the finished job reports the billed amount in usage.cost. How much does an AI video cost? sets it beside the other models.

Can I add sound to a Grok Imagine clip?

Not in the same request: the catalog lists no audio for this model. Add a voiceover or music afterwards, as in Add a voiceover to a video, or pick a model that generates audio with the clip (AI video with sound).