GPT-6 Sol & Claude Opus 5.5 are liveGPT-6 Sol at half the price of 5.6 Sol

Grok Imagine Video 1.5 API

AI Video

Grok Imagine Video 1.5 is xAI's first live video route on OmniaKey: it renders text-to-video and image-to-video clips through the same asynchronous media API, priced per second of output instead of per call.

Read the docsGet your API key

xAI

Grok Imagine Video 1.5

Mode
Prompt0 / 12000
Duration
Resolution

Estimated $1.50$0.30/s × 5s

Your result will appear here once you run the model.

01

Where it fits

Grok Imagine Video 1.5 is xAI's first live video route on OmniaKey: it renders text-to-video and image-to-video clips through the same asynchronous media API, priced per second of output instead of per call.

Pick Grok Imagine Video 1.5 when you need a short clip from a prompt or a single first frame and want the cost to scale with clip length; step up to the Seedance, HappyHorse, or Wan 3.0 routes when you need audio, longer clips, or higher fidelity.

02

What it does

Text to video

grok-imagine-video-1.5

Generate a short video clip from a text prompt.

Image to video

grok-imagine-video-1.5

Animate a still image into a short video clip.

03

In depth

Grok Imagine Video 1.5 is xAI's video model on OmniaKey, listed as grok-imagine-video-1.5 and sold at one flat rate per output second. It renders at up to 720p, takes durations from 4 to 15 seconds, and accepts nine aspect ratios, with 16:9 as the default.

The model is built around sound: clips arrive with a stereo track of effects, ambience and dialogue, and the prompt's sound sentence influences the picture as well. In our coffee clip the sound line mentioned a spoon, and a spoon and a hand duly appeared at 4.5 seconds.

xAI made 1.5 generally available on 16 June 2026, and the model has kept the same request shape since. Our recorded run is a single 6-second 720p text-to-video job from 24 October 2026, and it also produced the two error paths and the link expiry quoted below.

Sound-led social clips

Write a Sound line after the visual description and the model will stage the noise: a spoon on a saucer, footsteps behind a door. Short vertical spots that live on their audio track get most of their character from that one line.

Unusual frame shapes

Nine ratios include 21:9, 5:2, 3:2 and 2:3, which covers a banner strip or a print-style portrait without a reframe. Set ratio in the parameters; 16:9 is what you get if you leave it out.

Drafting a beat before committing

Because the price scales with seconds and a failed job is refunded, a 4-second test on the shortest allowed length is a cheap way to check a camera move before ordering the full 15 seconds.

04

Hands-on notes

  • Our 6-second 720p Grok Imagine Video 1.5 clip was ready 46 seconds after we sent the request and came back with an audio track.

    Read the test
  • Our 6-second Grok Imagine Video 1.5 request returned a file 6.04 seconds long and was billed for the 6 seconds we asked for.

    Read the test
  • In our coffee clip, Grok Imagine Video 1.5 put a spoon and a hand on screen, both prompted only by the sound line we wrote.

    Read the test
  • When we reused our job's idempotency key with a changed prompt on 24 October 2026, Grok Imagine Video 1.5 returned HTTP 409 instead of a new task.

    Read the test
  • Our Grok Imagine Video 1.5 download link expired exactly one hour after the job finished, so we polled the task again for a fresh URL.

    Read the test

05

Limits to plan around

  • 720p is the ceiling here. xAI's own API also lists 480p and 1080p, but this catalog serves the model at up to 720p, so a 1080p request has no tier to land on.
  • Durations below four seconds are refused even though xAI's own range starts at one second, so a two-second sting cannot be requested here.
  • The signed download link lasts about an hour while the task record lasts a day, and a missed window means polling the same id again rather than re-rendering.
  • A reused idempotency key with a changed body returns a conflict instead of a new job, which is easy to hit when a retry loop edits the prompt.

06

Compare nearby models

ModelBest forOperationsListed price
Grok Imagine Video 1.5This page
grok-imagine-video-1.5
One flat per-second rate, 720p, 4 to 15 seconds, nine ratios, sound by default.Text to video · Image to video
$0.30
per second
Kling 3.0
kling-3.0
Fixed 5, 10 or 15 second clips with a mode that sets the resolution.Text to video · Image to video
$0.80–$3.00
per call
MiniMax H3
minimax-h3
768P or 2K, reference media, and no way to silence the track in the request.Text to video · Image to video · Reference to video
$0.0269–$0.0431
per second

08

Call it

Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.

Create, poll, download
# 1. Create the job. 202 Accepted; the body carries the task id,
#    the pre-charged price, and a poll_url. Reuse one Idempotency-Key
#    across retries so a timeout never starts a second paid job.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: grok-imagine-video-1.5-001" \
  -d '{
    "model": "grok-imagine-video-1.5",
    "type": "video",
    "operation": "text_to_video",
    "input": {"prompt": "a paper boat on a quiet lake at sunset"},
    "parameters": {"duration": 6, "resolution": "720p"}
  }'

# 2. The same model id and the same poll loop serve the other operations:
#      image_to_video         media[] type "first_frame"
#    Reference media travel alongside the prompt:
#    {"prompt": "...", "media": [{"type": "first_frame", "url": "https://example.com/first-frame.png"}]}
#    A public HTTPS URL or a base64 data URI both work.

# 3. Poll GET /v1/media/tasks/{id} until status is
#    succeeded | failed | canceled. Progress carries percent and stage.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks/TASK_ID \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY"

# 4. Download assets[0].url before expires_at, or poll again for a
#    fresh signed URL. Failed and canceled jobs are refunded.
curl -sS -o out.mp4 "$(jq -r '.assets[0].url' result.json)"

09

Pricing

Video jobs are billed per second of rendered output, not per token. Failed or canceled jobs are refunded automatically.

Per second of output

$0.30per second

More models from xAI

Related articles

Frequently asked questions

How do I call Grok Imagine Video 1.5 on OmniaKey?

POST to /v1/media/tasks with the model id, type "video", and operation "text_to_video" — or "image_to_video" with a first frame in input.media[] with type "first_frame". Reuse one Idempotency-Key across retries so a timeout never starts a second paid job.

What does Grok Imagine Video 1.5 cost?

Grok Imagine Video 1.5 bills $0.30 per second of rendered video. Creation pre-charges the requested duration up front from your prepaid balance, and failed or canceled jobs are refunded automatically.

Why does the API return a task id instead of the video?

Video renders take from about a minute to several minutes, so OmniaKey runs every video route through one asynchronous lifecycle. Poll GET /v1/media/tasks/{id} until the status is succeeded, then read the signed asset URL.

What duration can I request?

Between 4 and 15 seconds, set with duration, and the value is required. xAI's own API accepts as little as one second, but this catalog's minimum for the model is four, so anything shorter comes back as a price error before a job exists.

Does Grok Imagine Video 1.5 make sound?

Yes, by default. Clips carry effects, ambience and dialogue generated with the picture, and our file had a stereo AAC track. xAI's API exposes a generate_audio false switch; with the track already downloaded, a copy with the video stream and no audio is one edit away.

Why did my retry return HTTP 409?

Because the idempotency key was reused with a different body. The key identifies one request, and changing the prompt while keeping the key makes the retry a conflict rather than a second job. Send a new key for a new prompt, and keep the old key only for replaying the identical request.

How long is the download link good for?

About an hour, and ours expired exactly on the hour after the job finished. The task record survives for a day, so fetch that same record again to get a fresh signed URL instead of paying for a second render of the same prompt.

Can it animate a still image?

The catalog lists text-to-video and image-to-video, and xAI's own API takes an image as the first frame with an optional prompt, keeping the picture's aspect ratio. Our recorded job was text-to-video, so run one short image job and confirm the request shape before wiring it into a pipeline.