Short answer
Grok Imagine video is xAI’s capability for generating short clips from text or images, with generated audio. The model information summarized here distinguishes grok-imagine-video-1.5, which lists native 1080p, from grok-imagine-video-1.5-lite, whose 1080p output is described as upscaled from 720p. Both are listed for clips up to 15 seconds. For selection, weigh native resolution and listed reference controls against output cost. The API flow is described as asynchronous: submit a job, poll its request ID, and retrieve the completed video. Exact request details still need confirmation in the current API reference.
What Grok Imagine video can generate
The model comparison describes text-to-video and image-to-video generation with audio. It lists reference-image, voice-reference, and pinned-frame controls for model 1.5, but does not establish exact maximum counts. Those controls are not listed for Lite in the comparison; that does not confirm they are unavailable.
| Requirement | grok-imagine-video-1.5 |
grok-imagine-video-1.5-lite |
|---|---|---|
| Resolution | 480p, 720p, or native 1080p | 480p, 720p, or 1080p upscaled from 720p |
| Duration | Up to 15 seconds | Up to 15 seconds |
| Audio | Native audio, including lip-synced speech | Native audio, including lip-synced speech |
| Reference and frame controls | Reference and pinned-frame controls are listed; exact counts are unconfirmed | Controls are not listed in the comparison; availability is unconfirmed |
These are API-model specifications, not proof that every consumer interface exposes every option. The model comparison distinguishes native 1080p on model 1.5 from upscaled 1080p on Lite. It does not specify frame rate, so third-party claims of 24 fps should be treated as unconfirmed for these API models.
Selection rule: Consider Lite if lower output cost matters more than native 1080p and the listed reference controls. Consider model 1.5 if native 1080p or those controls matter to your workflow. If your delivery requires a particular frame rate or precise control limits, confirm them before committing.
When it is—and is not—a good fit
Grok Imagine is worth evaluating when your brief calls for short clips, image animation, generated audio, or reference-guided scenes. For example, test whether animating a still product image produces a usable short social clip. These are selection examples, not guarantees of output quality.
Run a pilot or look elsewhere if you need clips longer than 15 seconds, a specific frame rate that is not listed, or visual consistency beyond the documented controls. Feature specifications alone do not establish output quality for your subject, brand, or motion. Test representative inputs at your intended delivery resolution, then assess motion, audio, reference fidelity, and editing effort.
Illustrative prompt to test: “A slow, steady camera push-in on a reusable water bottle on a sunlit desk; soft window light, gentle condensation, no other objects moving; quiet room ambience.” This is a sample prompt, not a tested Grok output.
Access and cost: subscription versus API
Consumer subscriptions and API usage are separate purchase routes. A subscription price is not an API rate, and consumer-plan access does not establish API eligibility or billing terms.
Direct xAI API pricing
The model pages summarized here list output prices using a dollar symbol per generated second, varying by model and resolution. The currency code for these displayed rates has not been confirmed. The Lite model information also says image or video inputs can incur charges. The available information does not establish whether that same input-billing condition applies to model 1.5. Verify the selected model’s current billing terms and the price shown for your account before use.
| API model | 480p | 720p | 1080p |
|---|---|---|---|
grok-imagine-video-1.5 |
$0.08/sec | $0.14/sec | $0.25/sec |
grok-imagine-video-1.5-lite |
$0.02/sec | $0.03/sec | $0.14/sec |
grok-imagine-video |
$0.05/sec | $0.07/sec | Not listed |
Worked example: A 10-second, 720p output would be $1.40 with model 1.5 or $0.30 with Lite, before any applicable input charges. A 10-second native 1080p output with model 1.5 would be $2.50. A 10-second Lite output at 1080p would be $1.40, but is described as upscaled rather than native 1080p. These calculations multiply the listed output rate by clip duration; they are not consumer-subscription prices, and the currency code is unconfirmed.
Consumer subscription access
The consumer pricing information summarized here displays Free at $0/month, Plus at $8/month, Premium at $30/month, and Super at $100/month. It promotes AI video creation for Premium and specifically advertises 1080p, 15-second videos for Super. The available information does not establish an exact video quota or confirm video eligibility for every listed plan.
The currency code and consumer checkout channel for these monthly figures have not been confirmed. They are reference amounts, not an established U.S. checkout quote. Confirm the plan, price, and usage limits applicable to your region and account before subscribing.
Practical choice: Use a consumer plan if its available interface features and usage limits fit your needs. Use the API when you need programmatic generation and can budget by model, resolution, duration, and any billable inputs. For an intermediary provider, check its model-selection controls, rates, limits, and terms separately; its endpoint and charges are not xAI’s direct API.
Direct xAI production API: request and polling flow
The xAI video API is described as asynchronous: a generation request returns a request_id, which can be polled until the job is done or failed; the completed result is read from video.url. The exact submission and polling paths, authentication scheme, and accepted request fields and values are not confirmed here. Do not treat the outline below as executable API instructions; check the current xAI REST reference for those details before implementation.
Conceptual flow only:
submit generation request using the documented endpoint and authentication
request_id = returned request ID
repeat:
result = poll request_id using the documented endpoint and authentication
until result status is done or failed
if result status is done:
retrieve result.video.url
else:
handle the failed job according to the documented error response
In an application, retain the request ID, show that generation is in progress, and handle completed and failed jobs. The API flow described here does not establish a webhook option or numeric request limit for model 1.5. A 10-requests-per-second limit is reported for Lite; confirm the applicable limit for your model and account before sizing production traffic.
Bottom line
Grok Imagine video may fit projects that need short clips and generated audio, provided its resolution and controls meet the brief. Consider Lite when lower output cost is the priority and upscaled 1080p is acceptable; consider model 1.5 when you need native 1080p or its listed reference and frame controls. Before production rollout, confirm the current endpoint, authentication, request schema, billing conditions, frame-rate requirements, and account-specific limits in xAI’s documentation.
Sources
- xAI video generation
- xAI video API REST reference
- Grok Imagine Video 1.5 model information
- Grok Imagine Video 1.5 Lite model information
- xAI pricing
Sources
Video Generation: Grok API Capabilities