Yes, Grok makes videos. Grok Imagine, xAI's image and video generator, produces short clips with synchronized sound inside the Grok app (iOS, Android and grok.com) and through the xAI API. The current model, Grok Imagine Video 1.5, generates clips of 1 to 15 seconds at 480p, 720p or 1080p with audio on by default, including lip-synced speech, and it accepts a starting image, reference images, reference voices and pinned keyframes. Video generation needs a paid Grok subscription in the app, or pay-per-second API access for developers. A Grok clip is a shot rather than a finished post, so a Short or YouTube video still needs a script, a voiceover, captions and an edit.
This guide is for creators and marketers who use Grok or are comparing it with Gemini, ChatGPT and Claude. It covers what Grok Imagine is, which plans include video, clip length and resolution, the controls you get, a step-by-step walkthrough, the API, watermarks and moderation, a workflow that turns clips into a full video, a comparison with the other assistants, costs and common mistakes. Facts are current as of October 2026; xAI changes plans often, so confirm limits inside the app.
Related guides: best Sora alternatives in 2026, can Gemini make videos, can ChatGPT make videos and can Claude make videos.
What Is Grok Imagine?
Grok Imagine is the creative mode of Grok, the assistant built by xAI. It started as an image generator and added video in August 2025, when it became one of the first chat assistants to generate clips with sound on a phone. In 2026 xAI released Grok Imagine Video 1.5, first as an API preview on June 3, 2026 (xAI) and then in the Grok app, raising the quality ceiling to native 1080p and adding reference images, reference voices and keyframes.
Three things distinguish Grok's generator from the others. It is fast, usually returning a clip in well under a minute. It is cheap per second compared with Google, Runway or Kling. And it lives inside a general assistant, so you can research a topic, write a script and generate the opening shot in one app. The trade-offs are weaker prompt adherence on complex scenes, a plan structure that has changed several times in 2026, and a moderation approach that has been looser than Google's, which drew criticism in 2025 and means platforms will apply their own rules to what you post.
Which Grok Plans Include Video Generation?
Video generation is a paid feature. Free Grok accounts lost access to Imagine video in March 2026, according to plan trackers, and xAI then introduced a cheaper entry tier: SuperGrok Lite at $10 a month, launched on March 25, 2026 with basic image and video generation (short 480p clips), sits below the standard SuperGrok plan at $30 a month or $300 a year, which includes 1080p video and longer clips, with SuperGrok Plus and SuperGrok Heavy above it for heavier usage and priority rendering (Tech Insider). X Premium subscribers have historically received Grok features inside X as well, with the top tier matching SuperGrok.
xAI does not publish a single table of clip length and resolution per plan, and the figures circulated by third parties conflict, so the reliable method is to open Imagine in the app, start a video and read the length and quality options your plan shows. If you need a guaranteed spec, the API is the only place xAI documents it precisely.
How Long Can Grok Videos Be? Length, Resolution and Audio
xAI's developer documentation is the clearest source. The grok-imagine-video-1.5 model generates clips from 1 to 15 seconds, with 8 seconds as the default. Output resolution is 480p, 720p or 1080p, and 1080p is rendered natively for text-to-video and image-to-video. The cheaper grok-imagine-video-1.5-lite model reaches 1080p by upscaling a 720p render, and the older grok-imagine-video model tops out at 720p (xAI docs).
Audio is included by default. Grok Imagine 1.5 generates sound effects, ambience and speech, and the speech is lip-synced to on-screen characters; you can request a silent clip instead. Reference-to-video on the 1.5 model can give up to three speakers preset voices from reference audio. In the app, an extend feature reported in August 2026 continues a clip to roughly 30 seconds in total, but a single generation stays within the 15-second ceiling.
Aspect ratios are generous for social creators: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9 and 5:2, with 16:9 as the default. Choose 9:16 before generating anything for Shorts, Reels or TikTok; image-to-video ignores the setting and matches the input image instead, so start from a vertical image for vertical output. Our video aspect ratios guide explains the sizes each platform wants.
What Can You Control in Grok Imagine?
Text-to-video: describe the subject, action, camera move, lighting and mood, and set the length and aspect ratio.
Image-to-video: upload a still or pick a generated image and describe the motion. The clip keeps the detail and lighting of the frame, which is the most reliable way to get a consistent look.
Reference images and voices: on the 1.5 model you can supply reference images for a character or product and reference audio for up to three speakers, so a face or a voice carries across clips.
Keyframes and last frame: the 1.5 model accepts up to four pinned keyframes with timestamps and a pinned final frame, which lets you plan a shot that starts on one image and ends on another.
Editing and extension: the classic grok-imagine-video model edits an existing clip (changing a background, for example) at up to 720p, keeping the original duration, which is capped at 8.7 seconds for edits.
What you can't control is as important. There is no timeline, no caption layer, no narration track beyond generated speech, and no multi-scene storyboard with guaranteed continuity. Each generation is one shot.
How to Make a Video With Grok: Step by Step
Step 1: open Grok on your phone or at grok.com on a paid plan and switch to Imagine, then choose Video.
Step 2: write the prompt as a shot description, not a story. A good pattern is subject, action, setting, camera, lighting, length: “A barista pours latte art in a sunlit café, slow push-in on the cup, warm morning light, 8 seconds.” Add “no text on screen” to avoid garbled lettering.
Step 3: set 9:16 for vertical platforms and the highest resolution your plan allows, then generate. Expect to run two or three variations; speed is Grok's advantage, so use it.
Step 4: for consistency, generate a still first, refine it until it's right, then animate it with image-to-video. For a recurring character, use reference images on every clip.
Step 5: download the MP4. Then move to an editor or a video tool, because the clip is one shot and the next section explains what it still needs.
Grok Imagine for Developers: The xAI API
Developers can call the same models through the xAI API with three options: grok-imagine-video-1.5 (native 1080p, references, keyframes), grok-imagine-video-1.5-lite (the lowest price per second, 1080p by upscaling) and grok-imagine-video (the classic 720p model, required for video editing and extension). Generation is asynchronous: you submit a request, poll for completion, and download the result promptly because output URLs are temporary (xAI docs).
Pricing is per second of output and depends on model and resolution. xAI's pricing page lists the current rates; third-party trackers in October 2026 quoted roughly $0.02 per second for the lite model at 480p and rising with resolution on the full 1.5 model, which makes Grok one of the cheapest hosted generators per second. For a scripted pipeline, that means a 15-second 1080p clip costs a few dollars at most. The API also returns a moderation flag with each video, so builders can filter outputs before they're shown to users.
Watermarks, Moderation and AI Labels
xAI's developer documentation does not describe a visible watermark or C2PA Content Credentials on Grok Imagine output, so don't count on the file to label itself, and don't assume platforms can't tell either. YouTube requires disclosure of realistic synthetic content, TikTok labels AI content and reads Content Credentials where present, and Meta applies labels to realistic generated media. Apply the label yourself when the footage is realistic; the rules by platform are in our AI content labels guide.
On moderation, Grok rejects some prompts and flags others, and the API reports whether a clip passed moderation. Grok's policies have been more permissive than Google's or OpenAI's were, which is part of why it attracted attention in 2025. That doesn't loosen the platforms' rules: a clip that Grok will generate can still be removed by YouTube or TikTok for violating their policies on violence, nudity or impersonation. Treat the strictest platform you post to as the standard.
Why a Grok Clip Isn't a Finished Video Yet
A 15-second clip with ambient sound is a shot. A Short that performs has a hook in the first second or two, a reason to keep watching, a payoff, a voice and captions for the majority of viewers watching on mute, and usually a visual change every two to four seconds. Even a 30-second extended Grok clip is one continuous scene.
Generated clips also drift between generations, so a face or a product changes slightly from shot to shot unless you anchor it with references; on-screen text often comes out garbled; and every retry spends your plan's usage. YouTube's monetization policies exclude mass-produced, repetitive content, so a channel of stitched clips without an original script is a weak strategy, as our guide to the YouTube AI content monetization policy explains. Use Grok as a footage source inside a normal production workflow.
Workflow: Turn Grok Clips Into a Full Short or YouTube Video
Step 1, script first: write the video in Grok, ChatGPT or Claude before generating anything. A 45-second Short is about 110 words at a natural pace; the free video length calculator converts word count to runtime. Our guides to video hooks that stop the scroll and TikTok script templates give you structures.
Step 2, decide which shots need generation: usually the hook and one or two moments that can't be filmed or found in stock. Generate those in Grok at 9:16 and download them.
Step 3, build the rest in a video tool. In Vidpal Studio, paste the script, pick a voice and a caption preset, let Studio match stock footage or AI images to each scene, and swap in your Grok clips where they belong. Studio's Hook Intro can also open with a Text → Video or Image → Video clip of its own, so you can compare Grok's opener with Studio's. A 9:16 Short costs 1 credit and a 16:9 video of 1 to 10 minutes costs 5; the Studio page shows the video sources and the pricing page lists plans from $19 a month.
Step 4, if you recorded yourself instead, upload the footage to an AI Editing project for silence and filler-word removal, captions, zooms and b-roll, and use the Grok clip as a cutaway. Our guide to adding b-roll automatically covers that step.
Step 5, publish and label: export, add the platform's AI disclosure where the Grok footage is realistic, and schedule it. On Pro and Business plans Vidpal publishes to YouTube, Instagram and TikTok on a schedule; the scheduler page explains how.
Grok vs Gemini vs ChatGPT vs Claude for Video
Grok: generates 1 to 15-second clips with lip-synced audio at up to 1080p, fast and cheap, on paid plans; no video analysis documented, no caption or narration layer.
Gemini: generates up to 10-second clips with audio using Gemini Omni on Google AI plans, extends to 40 seconds, adds Flow for multi-shot work, attaches SynthID watermarks, and analyzes uploaded and YouTube videos. The most complete package; see can Gemini make videos.
ChatGPT: no video generation since OpenAI shut Sora down in 2026; strong scripting and image generation, and it can drive a video tool through a connector. See can ChatGPT make videos.
Claude: no video generation or video input; the strongest writer and coder, with Claude Motion for code-based animations on Team and Enterprise plans and MCP connectors on every plan. See can Claude make videos.
Pick Grok for quick, cheap realistic clips and a permissive creative mode; Gemini for quality, controls and video analysis; ChatGPT or Claude for the script and for operating a tool that builds the whole video. Many creators use two of them. Our three-way comparison of ChatGPT vs Gemini vs Claude for video goes deeper on the assistants that don't generate footage.
What It Costs to Make Videos With Grok
Grok: SuperGrok Lite at $10 a month covers short, low-resolution clips; SuperGrok at $30 a month ($300 a year) adds 1080p and longer clips; higher tiers add volume and priority. Prices are US figures reported in September and October 2026 and change without much notice, so check the plan page in the app.
API: per second of output, from about $0.02 at 480p on the lite model; xAI's pricing page has the exact rates for each model and resolution.
Assembly: Vidpal's Free plan includes 4 one-time credits for exploring; Studio exports need a paid plan, from Starter at $19 a month for 15 credits to Pro at $39 (40 credits, publishing) and Business at $69 (100 credits).
For most creators the cheapest working setup is SuperGrok Lite or a free Gemini route for the occasional generated shot, plus a Starter plan to turn scripts into finished videos.
Common Mistakes When Making Videos With Grok
Writing a story instead of a shot: Grok renders one scene per generation. Describe a single moment with a camera move and a duration.
Generating landscape clips for vertical platforms: set 9:16 first, or start from a vertical image, because image-to-video ignores the aspect ratio setting.
Asking for on-screen text: lettering comes out garbled. Add titles and captions in the edit.
Posting a raw clip as a Short: no hook, no voice, no captions. Build the video around the clip.
Relying on Grok's moderation as a guide to platform rules: the platforms are stricter, and realistic footage still needs an AI label.
Trusting third-party plan tables: xAI's plans have changed several times in 2026. Read the limits inside the app before you commit to a workflow.
Key Takeaways
Takeaway one: Grok makes videos through Grok Imagine, with clips of 1 to 15 seconds, native audio and lip-synced speech, at up to 1080p on the 1.5 model.
Takeaway two: video generation needs a paid Grok plan (SuperGrok Lite from $10 a month) or the xAI API, which bills per second of output.
Takeaway three: Grok supports 9:16, reference images and voices, keyframes and image-to-video, but has no timeline, captions or narration layer.
Takeaway four: xAI documents no watermark on Grok videos, so label realistic clips yourself; platforms enforce their own content rules regardless of Grok's moderation.
Takeaway five: a Grok clip is one shot; a video creation tool such as Vidpal Studio turns the script and the clip into a finished, captioned Short.
Frequently Asked Questions
Can Grok make videos? Yes. Grok Imagine generates short video clips with sound inside the Grok app and through the xAI API. The Grok Imagine Video 1.5 model makes clips from 1 to 15 seconds at up to 1080p, with lip-synced speech, sound effects and support for reference images and voices.
Is Grok video generation free? No. Free Grok accounts don't include Imagine video as of 2026. The cheapest route is SuperGrok Lite at $10 a month for short low-resolution clips; SuperGrok at $30 a month adds 1080p and longer clips. Developers pay per second through the xAI API.
How long can a Grok video be? A single generation is 1 to 15 seconds on the 1.5 model, with 8 seconds as the default. An extend feature in the app continues a clip to roughly 30 seconds in total. Longer videos are edits that join several clips with narration and captions.
Does Grok Imagine make vertical videos for TikTok and Shorts? Yes. Grok supports 9:16 along with 1:1, 16:9, 4:3, 3:4, 3:2, 2:3, 21:9 and 5:2. Choose 9:16 before generating, or start image-to-video from a vertical image, because image-to-video matches the input image's shape.
Does Grok add a watermark to generated videos? xAI's developer documentation doesn't describe a visible watermark or Content Credentials on Grok Imagine output. Platforms may still detect or require labels for realistic AI footage, so apply YouTube's, TikTok's or Meta's AI disclosure yourself.
Can Grok watch or analyze a video? Grok's documented inputs are text, images and files rather than video uploads. For feedback on a video, give Grok the transcript or a few still frames. Gemini is the assistant that analyzes uploaded and YouTube videos directly.
Which is better for video, Grok or Gemini? Gemini has the more complete toolset: longer extended clips, Google Flow for multi-shot consistency, SynthID watermarking and video analysis. Grok is faster and cheaper per clip with looser moderation. For a finished Short, both need a script, narration, captions and an edit, which a tool like Vidpal Studio supplies.
Start Making Videos With Grok
Grok is a quick, inexpensive source of generated shots, and in 2026 that's a genuinely useful thing to have inside a chat app. It isn't a video editor, a narrator or a publisher, so plan the script first, generate only the shots that need it, and build the video around them.
To turn a script and a clip into a finished, captioned Short, start in Vidpal Studio or connect Vidpal to ChatGPT or Claude. You can create a free account to explore, then pick a plan when you're ready to export.
