What Is Grok Imagine and How Do You Make a Video With It? (2026)

Grok Imagine is xAI's image and video generation model: it turns a text prompt (or a starting image) into a short clip with motion and generated sound, and its current video version is Imagine Video 1.5, released on June 16, 2026. The quickest way to try it without an X or Grok subscription is a browser generator – GenModelHub's Grok video generator runs the Grok Imagine video model from a text prompt, with no app to install. GenModelHub is not affiliated with xAI. Updated October 18, 2026.

How to make a Grok Imagine video with GenModelHub (3 steps)

This is the short route. It runs in a phone or desktop browser and the generator is already set to Grok Imagine.

Describe the scene. Open the Grok video generator and write one or two sentences: who or what is in it, what happens, where, and how the camera moves. Example: "A red fox trots across a snowy field at sunrise, low tracking shot, breath visible in the cold air."
Pick length and format. Choose 6 or 10 seconds, then a shape: 9:16 for TikTok, Reels and Shorts, 16:9 for YouTube, 1:1 or 2:3 for feeds, or 3:2.
Press Generate and download. A clip usually takes one to three minutes. Play it, download the 720p MP4 (sound included), or change one part of the prompt and run it again.

How to write the prompt (there is no photo to choose)

On GenModelHub, Grok Imagine is text-to-video: you do not upload a photo. What you write is the whole input, so it matters more than anything else.

Lead with subject and action. "A skateboarder lands a kickflip" beats "an epic, cinematic, ultra-detailed scene".
Add place and light. "Empty parking lot at dusk", "warm kitchen light".
Give the camera one instruction. "Slow push-in", "low tracking shot", "wide shot". Two or three camera moves in one 6-second clip tend to fight each other.
Name the sound if it matters. "Waves and gulls", "crowd cheering", "sound of rain". Sound is generated with the video.
Keep one idea per clip. For a series, reuse the same description of your character and change only the action.
Avoid: real people without their consent, sexual content, and anything illegal. If the model provider rejects a prompt, the generation fails and your credits are refunded automatically.

A good prompt: "Close-up of hands kneading bread dough on a floured wooden table, warm kitchen light, slow push-in, soft sound of dough being pressed." A weak prompt: "cool video of food, amazing, 4K, best quality."

To animate a photo you already have, use the image-to-video page instead.

What you get

A 6- or 10-second 720p MP4 with sound, in 16:9, 9:16, 1:1, 2:3 or 3:2, ready to post. Generation takes about one to three minutes.

Price: GenModelHub sells credits one time, with no subscription: $8.99 buys 30 credits and $24.99 buys 90. Grok Imagine is billed at 1 credit per second, the lowest tier on the site, so a 6-second test is 6 credits and a 10-second clip is 10. The cheap AI video generator page lists every model's cost side by side. If a generation fails, the credits are returned automatically.

Not happy with the result? Change one thing in the prompt and run it again. Test a new idea at 6 seconds before you spend 10.

If the result doesn't look right

Problem	Fix
The motion is jittery or the scene changes mid-clip	Describe one action and one camera move, not a sequence
People look inconsistent	Describe the person once, in the same words, and keep the shot simple (one subject)
Text on signs or shirts is garbled	Remove written text from the prompt; AI video models still struggle with lettering
The sound is wrong or missing the thing you wanted	Name it explicitly ("rain on a tin roof, distant thunder")
Wrong shape for the platform	Re-run in 9:16 for TikTok, Reels and Shorts or 16:9 for YouTube
Generation failed or was rejected	Credits come back automatically; remove any sexual, real-person or unsafe wording and retry
You need a clip longer than 10 seconds	Generate several clips with the same character description and join them in any video editor

What Grok Imagine is, in facts

Maker: xAI, the company behind the Grok assistant. Grok Imagine covers both images and video.
Video API launch: xAI's API release notes list video generation and a revamped image generation as released on January 28, 2026.
Imagine 1.0: launched on February 2, 2026 with clips up to 10 seconds at 720p and audio generated at the same time, according to Gigazine.
Imagine Video 1.5 Preview: on June 3, 2026 xAI released grok-imagine-video-1.5-preview, "our latest image-to-video model," through the API, per xAI's announcement.
Imagine Video 1.5: announced June 16, 2026, as "improved quality at even faster speeds." It generates sound effects, ambience and dialogue alongside the video; xAI says "Speech is clearer and better synced" (xAI). A faster "Video 1.5 Fast" variant runs on grok.com/imagine and in the iOS and Android apps.
Imagine Video 1.5 with References: on July 31, 2026 xAI added image and voice references, up to seven references per generation, text-to-video ("Describe the shot — no starting image needed"), and native 1080p (xAI).
Newest video version: 1.5 is the newest Grok Imagine video model on xAI's news page as of October 18, 2026; the newest Imagine post there is Imagine Image 2.0 (August 7, 2026), an image model. We found no Video 2.0.
Length and resolution: xAI's Grok page describes video generation as up to 15 seconds at 720p. Third-party coverage (ExplainX) also gives 15 seconds for 1.5; GenModelHub offers 6 or 10 seconds at 720p.
Developer price: xAI's API page lists grok-imagine-video-1.5 from $0.08 per second and grok-imagine-video-1.5-lite from $0.02 per second (checked October 18, 2026; the page shows starting prices only).

How to use Grok Imagine on grok.com or in X

The official route is xAI's own apps. The details change often, so check the app for current limits.

Open grok.com/imagine, or the Grok app on iOS or Android, and sign in. xAI describes Grok as "free to try on the web and in the apps" on its Grok page, but that page does not say which Imagine features are limited to paid plans.
Choose video, write a prompt (or add a starting image), and generate.
Higher tiers add capacity: the July 31, 2026 release gave image and voice references first to SuperGrok Heavy and SuperGrok Plus subscribers in the US, then rolled them out to all tiers; text-to-video and native 1080p were generally available on grok.com/imagine, iOS and Android.

Subscription prices. We could not load xAI's own plans page (grok.com/plans) on October 18, 2026, so the numbers below are third-party and should be confirmed at checkout. One comparison site that read the plans page on October 3, 2026 lists SuperGrok at $30 per month, SuperGrok Plus at $100, SuperGrok Heavy at $300 and X Premium+ at $40, and says SuperGrok includes Imagine image and video. The same page lists a SuperGrok Lite at $10 per month; other posts we found do not mention it, so treat that one as unconfirmed. Posts on X about the launch of SuperGrok Plus also give $100 per month and mention 1080p video.

The cost of that route: a monthly plan, an X or Grok account, and plan-dependent limits. The benefit is the full feature set: image and voice references, 1080p and longer clips.

Comparison

	grok.com / X (xAI apps)	GenModelHub Grok video generator	xAI API	Veo 3 (Google)
Input	Text, image, references (by plan)	Text prompt	Text, image, references	Text or image
Result	Up to 15 s, 720p–1080p, with sound	6 or 10 s, 720p, with sound	Up to 1080p, with sound	Clips with sound (about 8 s)
Time	Seconds to minutes	About 1–3 minutes	Depends on your code	Minutes
Cost	Plan-based (third-party prices: $30–$300 per month)	1 credit per second; credits from $8.99 for 30 or $24.99 for 90, one-time	From $0.02–$0.08 per second	Subscription at launch: $249.99 per month
Limits	Account needed, plan-dependent caps	Text-to-video only, no photo upload for this model	Needs code and an API key	Subscription; availability changed since launch

Pick the xAI apps if you want references, 1080p or longer clips and already pay for SuperGrok. Pick the browser generator if you want one clip, paid once, with no account on X or Grok.

Tips that work on any tool

Short and concrete wins. One subject, one action, one camera move.
Iterate in small steps. Change one phrase per attempt so you learn what caused the difference.
Test short, finish long. Try ideas at 6 seconds.
Label it. Turn on the AI-generated label when you post to TikTok, Instagram or YouTube.
Consent first. Do not generate real people without their agreement.

FAQ

What is Grok Imagine? Grok Imagine is xAI's AI image and video generator, available on grok.com/imagine, in the Grok apps and through the xAI API. Its newest video model is Imagine Video 1.5 (June 16, 2026), updated on July 31, 2026 with references and native 1080p.

How do I make a video with Grok Imagine? Write a prompt on grok.com/imagine and generate, or use a browser generator: on GenModelHub, open the Grok video generator, describe the scene, pick 6 or 10 seconds and a format, and download the MP4 after one to three minutes.

How long are Grok Imagine videos? xAI's Grok page says video generation runs up to 15 seconds at 720p; Imagine 1.0 launched with 10-second clips. On GenModelHub, Grok Imagine clips are 6 or 10 seconds at 720p, with sound.

How much does Grok Imagine cost? In xAI's apps it depends on your plan (third-party sources list SuperGrok at $30 per month; confirm on xAI's site). Through xAI's API it is from $0.08 per second for Video 1.5. On GenModelHub it is 1 credit per second (a 10-second clip = 10 credits), with credits sold one time: $8.99 for 30 or $24.99 for 90, no subscription.

Is GenModelHub the official Grok website? No. GenModelHub is an independent service, not affiliated with xAI or X. Grok and Grok Imagine are names of xAI products, used here only to describe the model.

More: animate a photo with image-to-video and compare every model's cost.