Grok Imagine is the name xAI uses for image and video generation capabilities, with separate API model IDs for different tasks—not one interchangeable model. The documented image model is grok-imagine-image-2.0; video options include grok-imagine-video-1.5 and grok-imagine-video-1.5-lite. Available operations and controls vary by model, so choose based on your workflow’s inputs, outputs, and quality requirements. (xAI Imagine overview)
What can Grok Imagine do?
The xAI API documentation describes these capabilities:
- Generate images from text:
grok-imagine-image-2.0supports image generation, with controls including output count, aspect ratio, resolution, and response format. A request can generate up to 10 images. (Image generation docs) - Edit images: The Imagine overview describes natural-language editing, including edits that combine up to five reference images. (xAI Imagine overview)
- Generate video from text or a still image: Video 1.5 and Video 1.5 Lite support text-to-video and image-to-video, with clips up to 15 seconds. (Video overview)
- Generate video with audio: The overview lists native audio with lip-synced speech for both Video 1.5 and Lite. (Video overview)
- Use advanced video references and frame controls: Video 1.5 supports up to 14 reference images, up to three voice references, and pinned frames. Its listed output duration is 1–15 seconds, and it supports native 1080p. (Video overview)
These API specifications do not mean the same features or controls are available in the Grok consumer app or through third-party services. Check the documentation for the specific model and channel you intend to use.
Is there a production API?
Yes. xAI documents an Imagine API for image and video generation and editing. It specifies distinct model IDs and operations; API billing is separate from a Grok consumer subscription or a third-party service’s plan. (xAI Imagine overview)
Video generation is asynchronous: submit a request to POST /v1/videos/generations, use its returned request_id to poll GET /v1/videos/{request_id}, and retrieve the asset from video.url when the response reports status: "done". (Video REST reference; video extension documentation)
The documentation excerpts available for this guide do not establish all details needed for a copy-paste REST example, including the exact minimum payload and authentication headers. Confirm those in the current endpoint reference before implementing a request. This guide therefore describes the documented workflow without presenting an unverified command or response schema. It also does not assume that a particular SDK automatically polls for completion.
Which model should you choose?
| Requirement | Candidate | Documented distinction |
|---|---|---|
| Text-to-image generation or image editing | grok-imagine-image-2.0 |
Image generation plus natural-language editing; multi-image edits support up to five source images. (xAI Imagine overview) |
| Simple text-to-video or image-to-video when per-second cost matters | grok-imagine-video-1.5-lite |
Supports both input modes and lip-synced speech. Its 1080p output is upscaled from 720p, not native 1080p. (Video overview) |
| Native 1080p, reference images, voice references, or pinned frames | grok-imagine-video-1.5 |
Supports native 1080p, up to 14 reference images, up to three voice references, and pinned frames. (Video overview) |
The overview does not list reference-image, voice-reference, or pinned-frame controls for Lite; don’t assume they are supported there. Choose Video 1.5 if those documented controls or native 1080p are requirements. For simpler generation, Lite is xAI’s lower-priced-per-second option. These are capability and price distinctions, not evidence that one model produces better results for every prompt. (Video overview)
Editing and extension: The available excerpts do not establish which exact model ID supports video editing or extension. Check the operation-specific xAI documentation before selecting a model for either task; don’t infer that a similarly named model supports it.
API pricing: compare by billing unit
xAI states that prices are in USD. Image generation is priced per output image; video generation is priced per generated second, with rates varying by resolution. Image inputs are charged separately where listed. (xAI API pricing)
- Image 2.0 output: Low quality is $0.04 per image at 1K, $0.05 at 1.5K, and $0.06 at 2K. Medium quality is $0.06 at 1K, $0.07 at 1.5K, and $0.08 at 2K. Image input is listed at $0.01 per image, so an edit using an input image can incur both input and output charges. (Image 2.0 pricing)
- Video 1.5 output: $0.08 per second at 480p, $0.14 at 720p, and $0.25 at 1080p. Image input is listed at $0.01 per image. (Video 1.5 pricing)
- Video 1.5 Lite output: $0.02 per second at 480p, $0.03 at 720p, and $0.14 at 1080p. Image input is listed at $0.01 per image. (Video 1.5 Lite pricing)
Worked examples using the listed output rates: A six-second 720p clip costs $0.18 with Lite (6 × $0.03) or $0.84 with Video 1.5 (6 × $0.14). A 10-second 1080p clip costs $1.40 with Lite (10 × $0.14) or $2.50 with Video 1.5 (10 × $0.25). These calculations exclude any applicable image-input charge. Check current pricing and the inputs used for your operation before budgeting. (xAI API pricing)
Production selection checklist
Before choosing a model for production, compare your requirements with its documented operation and pricing:
- Modality and inputs: Do you need image generation, image editing, text-to-video, image-to-video, reference-to-video, or video editing and extension? Confirm the operation is listed for the exact model.
- Video controls: Check resolution, duration, reference support, voices, and frame controls. If native 1080p matters, note that Lite reaches 1080p by upscaling 720p.
- Cost per usable result: Estimate output images or video seconds at the required resolution. Add image-input costs when you supply images, and account for retries in your own evaluation.
- API workflow: Video jobs require polling unless you implement another documented handling method. Plan for pending or failed jobs and download generated assets from temporary URLs promptly.
- Quality and reliability: Test representative prompts and inputs in your own workflow. The cited documentation describes features and rates, not a comparative quality ranking against other models.
- Terms and access channel: Confirm the current terms for your intended use, and distinguish direct xAI API access from the consumer app or a third-party hosted service.
Bottom line: xAI documents Grok Imagine API options for image generation, image editing, and video tasks. Use Image 2.0 for image operations; choose Lite for the lower listed video-generation rates on straightforward jobs, or Video 1.5 when its documented reference controls or native 1080p are needed. Verify the operation-specific model for editing or extension, and validate output quality and total cost on representative requests before committing to production.
Sources
Imagine Overview — xAI API Documentation
Image Generation — xAI API Documentation
Video Overview: Grok Imagine Video API — xAI API Documentation
Grok Imagine Image 2.0 — xAI API Documentation
Grok Imagine Video 1.5 — xAI API Documentation
Grok Imagine Video 1.5 Lite — xAI API Documentation
API Pricing — xAI API Documentation