Grok Imagine
Turn text or up to 7 reference images into short video with native audio, from xAI's Grok Imagine.
Made with Grok Imagine
Sample outputs. Open in Studio to generate your own.
Where Grok Imagine shines
Text To Video
Write a prompt and Grok Imagine Video renders a short clip with synchronized native audio.
Image To Video
Add up to seven reference images to guide motion and style in an image-to-video result.
Social Clips
Frame in portrait, landscape, or square at 480p or 720p for vertical and horizontal feeds.
Strengths
- Handles both text-to-video and image-to-video in one pipeline
- Accepts up to 7 optional reference images to steer image-to-video
- Generates synchronized native audio alongside the video in a single pass
- Three modes (Fun, Normal, Spicy) for different content styles
- Flexible 6 to 30 second duration in one-second steps
- Five aspect ratios spanning portrait, landscape, and square
Trade-offs
- Resolution is capped at 720p, with only 480p and 720p tiers offered
- Priced per second, so cost scales with duration and the 720p tier costs roughly 1.9x the 480p tier per second
- A text prompt is always required, while reference images are optional
At a glance
About Grok Imagine
Grok Imagine Video is xAI's video generation model, available on HexGen as a video pipeline. Start from a written prompt for text-to-video, or add up to seven reference images to drive an image-to-video result. The underlying xAI model generates synchronized native audio in a single pass, so your clip arrives with matching sound rather than a silent render.
Pick the look that fits your project with three generation modes, Fun, Normal, and Spicy, and frame it in any of five aspect ratios covering portrait, landscape, and square (2:3, 3:2, 1:1, 16:9, 9:16). Choose 480p or 720p, then set your clip length anywhere from 6 to 30 seconds in one-second steps.
On HexGen, pricing is per second of video and tiered by resolution: a per-second base for each resolution multiplied by the duration you select. That keeps short clips affordable and lets you scale up length and resolution only when a project needs it.
Starting points
Copy, tweak, and run. Good prompts get you most of the way there.
A neon-lit city street at night in the rain, camera slowly tracking forward past glowing signs, 16:9, 720p, 10 seconds.
Animate this reference photo of a coffee cup: steam rising and gentle light shifting across the table, 1:1, 6 seconds.
A golden retriever running through a sunlit meadow in slow motion, soft ambient nature sound, 9:16, 480p, 8 seconds.
Pay only for what you render. 1 USD = 1,000 HGcoins. HGcoins never expire and failed runs refund automatically.
Grok Imagine vs other models
Grok Imagine Video stands out for shipping synchronized native audio and offering Fun, Normal, and Spicy modes. Here is how it sits next to other video models in the catalog.
| Model | Quality | Speed | Cost | Choose it when |
|---|---|---|---|---|
Grok Imagine This xAI | Great | Fast | Lower cost | Choose this when you want video with native audio, three content modes, and image-to-video from up to seven reference images, at a budget-friendly per-second 480p or 720p price. |
Kuaishou | Best | Fast | Mid cost | Pick Kling 2-6 when you want a higher-fidelity video render and do not need Grok's Fun/Normal/Spicy modes. |
ByteDance | Best | Fast | Higher cost | Pick Seedance 2-0 for premium video quality when budget is less of a concern. |
Related models
Frequently asked questions
It is xAI's video model that turns a text prompt, optionally guided by reference images, into a short video clip with synchronized native audio.