Grok Imagine

Grok Imagine in short

Grok Imagine is xAI's generator for pictures and short videos. You describe a scene, or give it a photo to start from, and it produces images or a clip with sound, including spoken dialogue on the newest video model. It is part of the official Grok apps, grok.com and the xAI API.

From first launch to Video 1.5

Image generation has been part of Grok since August 2024, when Grok-2 started drawing pictures inside X. An in-house image model named Aurora followed in December 2024. The Grok Imagine name arrived on August 4, 2025, when xAI rolled out a new image-and-video tool to SuperGrok and X Premium+ subscribers in the Grok iPhone app. It could turn a text prompt or a still image into a short animated clip with generated audio, and it soon came to Android and the web.

In 2026 the tool kept growing. A revamped image generator and video generation reached the API in January. Grok Imagine Video 1.5 became generally available on June 16, 2026, with smoother motion, more believable physics and sound, effects and speech generated in the same pass as the picture. A faster variant, Video 1.5 Fast, went into grok.com/imagine and the apps. On the image side, Imagine Image 2.0 is now the model xAI recommends for pictures.

What Grok Imagine can make

Images

• pictures from a text description, up to ten variations per request through the API;
• edits of an existing photo by describing the change, with up to five source images per edit;
• many shapes, from square to phone-screen tall and cinematic wide formats such as 21:9;
• standard or higher resolution (1k or 2k) and a choice of quality level.

Videos

According to xAI's video documentation, the current Video 1.5 model can:

• create clips from 1 to 15 seconds, at up to native 1080p;
• start from text, from a still image, or from reference pictures of characters, products and places (up to 14 per request);
• generate lip-synced speech with up to three preset voices;
• pin the first frame, the last frame and a few key frames in between, so you control where a shot begins and ends.

An older video model is still offered for editing an existing clip or extending it by a few seconds, and a cheaper Video 1.5 Lite handles simple text-to-video and image-to-video.

Where you can use Grok Imagine

• grok.com/imagine and the Grok apps for iPhone and Android, with projects, a searchable library and several generations running at once.
• Inside X through Grok, with tighter rules on what can be generated in public replies.
• The xAI API for developers, with dedicated image and video endpoints. Video jobs run in the background: you submit a request, check back for the result and download the clip.

Access depends on your plan. In xAI's October 2026 X Subscription, the $8 Plus plan lists Imagine among its features and the $30 Premium plan adds more usage and is described as the tier for making AI images and videos (plan prices as of October 2026). Free access to Imagine has been offered and withdrawn more than once, so check the current offer on xAI's pricing page.

API models and prices

These are the Imagine models on xAI's models and pricing page (xAI list price, Oct 2026):

The image endpoint follows the OpenAI images format, so the OpenAI SDK works with xAI's base URL. Our Grok API guide covers keys and setup for the text models.

Rules, safety and controversy

Grok's image features have been among its most criticized. When Imagine launched in 2025, reviewers noted a "spicy" mode that allowed some explicit content. From December 2025, people used Grok's image editing to create sexualized pictures of real women and children from photos, which led to lawsuits, regulators' attention around the world and, in January 2026, a restriction of image generation in X replies to paying subscribers.

If you use Imagine, follow xAI's acceptable use policy and the law where you live. Do not create images of real people without their consent, never create sexual content involving minors, and label AI-made media when you share it, especially anything that could pass for a real photo or news footage.

Grok Imagine and Grok.Online

Our site is independent and focuses on conversation. The three Grok models in our chat can describe a photo, read a chart or help you write a detailed prompt for an image tool, but they cannot draw anything themselves. That is a deliberate choice: Imagine is a paid xAI product, and we only offer what we can give you for free. To learn more about the models we do run, see What is Grok?.

FAQ

Can I use Grok Imagine on Grok.Online?

No. Grok Imagine is not in our chat. Our models read images you attach but do not generate images or videos.

Is Grok Imagine free?

Mostly not. xAI lists Imagine on its paid plans, and free access has changed several times. Check xAI's pricing page for the current offer.

How long can Grok Imagine videos be?

Up to 15 seconds per clip on the current Video 1.5 model, according to xAI's API documentation.

Does Grok Imagine make sound?

Yes. Video 1.5 generates sound effects, ambience and lip-synced speech together with the picture.

Who owns Grok Imagine?

xAI, the company now part of SpaceX and branded SpaceXAI / SpaceXSI.