xAI Grok Imagine · 3 to 30 seconds
Grok Imagine Video Generator: Version 1.5 and the Original
Run the xAI Grok Imagine models in your browser and turn a still photo or a typed prompt into a silent clip. Version 1.5 covers 3 to 15 seconds; the original model stretches a single take to 30.
Sign in to use these settings, with 30 free credits · Grok Imagine 1.5 · 480p · 5s
Grok Imagine 1.5 vs the Original Grok Imagine
Two generations of the model are available on ImageVideoAI, and they share almost every setting. The real split is clip length: the newer one is tuned for short takes, the first release for long ones.
| Version 1.5 | Original model | |
|---|---|---|
| Clip length | 3 to 15 s, 6 s by default | 6 to 30 s |
| Resolution | 480p, 720p, 1080p | 480p, 720p, 1080p |
| Starting point | Uploaded image or typed scene | Uploaded image or typed scene |
| Aspect ratios | 16:9, 9:16, 1:1, 3:2, 2:3 | 16:9, 9:16, 1:1, 3:2, 2:3 |
| Audio | None | None |
| Closing frame | Not supported | Not supported |
Version 1.5 is the default on this page. Pick it for social clips, product loops and anything under 15 seconds. Switch to the original when a scene needs room to unfold, such as a slow walk through a street or a long drone-style move over a landscape, since it is the only option in the family that reaches 30 seconds.
A note on 1080p: both versions produce it by rendering at 720p and then upscaling the result. You get a file with 1080p dimensions, but not native 1080p detail, so weigh the extra credits against where the clip will be shown.
What the Grok Video Generator Is Good At
Judged by its settings on this site, the Grok video generator has a few clear strengths:
- Five aspect ratios. Besides the usual 16:9, 9:16 and 1:1, it renders 3:2 and 2:3, which suit photo-style frames and print-like portraits.
- Two ways in. Start from an uploaded picture, or skip the upload and describe the whole scene in words.
- Long single takes. The original model goes up to half a minute without a cut.
- A cheap preview tier. At 480p, a 6-second 1.5 clip costs 30 credits, handy for checking motion before paying for a sharper render.
There are also things it does not do. Clips come out without sound, so plan to add music or voice in an editor. You cannot pin a final frame, so the ending has to be described in the prompt.
Grok Credits Per Clip
Resolution and duration set the price of every render, and the tool shows the exact figure before you confirm. A failed job puts its credits back in your balance.
| Model | Setting | Credits |
|---|---|---|
| Version 1.5 | 480p · 6 s | 30 |
| Version 1.5 | 720p · 6 s | 57 |
| Version 1.5 | 1080p · 6 s | 113 |
| Original model | 720p · 30 s | 282 |
New free accounts get 30 credits valid for 30 days, which is exactly one 6-second 1.5 clip at 480p, plus 5 credits a day that reset rather than pile up. Free renders are watermarked, top out at 720p and stop at 10 seconds. Upscaled 1080p, clips past 10 seconds, watermark-free files and commercial use come with a paid plan or credit pack, all listed on the pricing page.
Prompt Tips for Grok AI Video
These habits help the model follow what you have in mind:
- Anchor the subject in the photo. Refer to it the way it looks, such as "the cyclist in the red jacket", so the motion stays on the right figure.
- Pick the ratio before you write. A 9:16 frame leaves room above and below the subject; a 3:2 frame favors side-to-side movement. Describe action that fits the shape.
- Write the ending into the text. With no closing-frame upload, the last moment of the clip comes only from your words, so state where things should finish.
- Pace long takes in stages. On a 20- or 30-second render, split the action into an opening, a middle and a close, and keep the camera to one slow move.
- Skip sound words. The output is silent, so describing rain noise or music only spends prompt space. Describe what the rain looks like instead.
- Draft at 480p. Check the motion on the cheap tier, then render the keeper at 720p or higher.
Three prompts suited to this model:
The cyclist in the red jacket pedals out of the frame on the left, autumn leaves lift off the road behind the wheels, the camera stays at street level, low golden light.
Wide 3:2 shot of a mountain lake at dawn. Mist drifts across the water from right to left, the sun clears the ridge, and by the end the whole surface glows orange. Slow, steady pan.
Vertical 9:16 portrait of a dancer on a rooftop. She turns once, her skirt flares, then she stops and looks straight at the camera as city lights blink on behind her.
Where Grok Clips Fit Best
A Grok Imagine video works best where silent motion is enough or where sound gets added later:
- Vertical social posts. Render at 9:16 and add a trending track in your editing app.
- Product loops for a store page. A 6-second 1:1 clip of a bottle turning on a plinth needs no audio.
- Background footage. Thirty-second takes on the original model make calm backdrops for slides, livestream intros or website headers.
- Ideas without a picture. When there is no image to start from, write the scene out and the model builds it from text alone. Our text to video AI guide goes deeper on prompt-only clips.
- Photo-shaped frames. The 3:2 and 2:3 ratios keep the proportions of a camera photo, so an animated print or poster crops less.
Grok Imagine vs Wan vs Vidu
Two families on this site sit closest to Grok: Wan matches its longest take, and Vidu adds closing-frame control. All of them are listed in the model menu of our image to video AI tool, so a single photo can go through each one.
| Grok | Wan | Vidu | |
|---|---|---|---|
| Longest clip | 30 s (original) | 30 s | 16 s |
| Top resolution | 1080p, upscaled | 1080p | 1080p |
| Sound | None | On by default | None |
| Last frame | No | Yes | Yes |
| Example price | 57 credits for 720p, 6 s | 59 credits for 720p, 5 s | 28 credits for 720p, 5 s (Q3 Turbo) |
Alibaba's Wan 3.0 matches the 30-second ceiling and adds sound plus first and last frame control, so it is the stronger pick when a long take needs audio or a fixed ending. Its full settings are on the Wan AI video generator page.
Vidu is silent as well, tops out at 16 seconds, and accepts both an opening and a closing image on its Pro and Turbo variants. Turbo is the low-cost option at 18 credits for a 540p, 5-second clip. The Vidu AI page compares both variants.
Against Vidu, which offers 16:9, 9:16 and 1:1, Grok adds the 3:2 and 2:3 shapes, and the original model runs to 30 seconds against Vidu's 16.
Grok Imagine FAQ
Can I use Grok Imagine for free?
With limits. Each new free account opens with 30 credits, enough for one watermarked 6-second clip from version 1.5 at 480p, and gets 5 credits each day, which on their own don't cover a clip. Free renders stay at or below 720p and 10 seconds. Paid plans and credit packs unlock upscaled 1080p, longer clips and files without a watermark.
What is the difference between Grok Imagine 1.5 and the original model?
Mostly length. Version 1.5 makes clips from 3 to 15 seconds, defaulting to 6, while the original runs from 6 to 30 seconds. Resolutions, aspect ratios and input types are the same on both, and neither produces sound or takes a closing frame. Use 1.5 for short clips and the original for long single takes.
Does Grok AI video have sound?
Not on this site. Both versions return silent clips, and the tool has no audio switch for them. If your scene needs ambience, voice or music, add it in an editor afterward, or choose a family that generates sound with the picture, such as Wan 3.0, Kling 3.0 or MiniMax H3 on the Hailuo page.
Is the 1080p option native resolution?
No. The 1080p setting renders the clip at 720p and then upscales it, so the file has 1080p dimensions without native 1080p detail. It costs 113 credits for 6 seconds on version 1.5, against 57 at 720p, and it is reserved for paid accounts. For previews, 480p at 30 credits is the cheaper route.
Can the Grok video generator make a clip from text only?
Yes. Both versions support text to video as well as image to video, so you can leave the upload empty and describe the full scene, subject and camera move in the prompt. The same lengths, resolutions and five aspect ratios apply, and the credit price appears before you start the render.
Compare other models
Image to Video AI
Turn any photo into a short AI video. Upload an image, describe the motion and pick a model, from Kling and Veo to Grok Imagine and Vidu. Start free with 30 credits.
AI Video Generator
An AI video generator with Kling, Veo, Seedance, Hailuo, Wan, Grok Imagine, Vidu and PixVerse. Create videos from text or images and pay only for what you use.
Vidu AI Video Generator
Use Vidu AI online: Vidu Q3 Pro and Q3 Turbo turn a photo or prompt into a 3 to 16 second clip with first and last frame control. See credits per clip.
Kling AI Video Generator
Use Kling AI to turn images and prompts into cinematic videos. Compare Kling 4.0, 3.0, 2.6 and 2.5 Turbo, read prompt tips and try Kling image to video online.
A Grok Imagine Alternative That Still Runs Grok
Run Grok Imagine on ImageVideoAI next to Kling, Veo, Vidu and more, and compare it with xAI's own API on length, resolution, sound and price, Oct 2026.
Turn One Image Into a Grok Clip
Upload a picture or write a scene, choose version 1.5 or the original, then set the ratio and length. The exact credit cost sits on the button until you press it.