Text-to-image generation
Turn a detailed prompt into posters, product scenes, social creative, concept art, and other production-ready visuals. Describe the subject, composition, lighting, style, and any text that should appear.
Generate and edit images with Grok Imagine Image 2.0 by xAI. Use up to 3 references, 1K or 2K output, two quality tiers, and 14 aspect options via API.
| Specification | Credits | Price (USD) |
|---|---|---|
| Model & ModalityGrok Imagine Image 2.01K · Low | 3.8per image | $0.038per image |
| Model & ModalityGrok Imagine Image 2.02K · Low | 5.7per image | $0.057per image |
| Model & ModalityGrok Imagine Image 2.01K · Medium | 5.7per image | $0.057per image |
| Model & ModalityGrok Imagine Image 2.02K · Medium | 7.6per image | $0.076per image |
| Model & ModalityGrok Imagine 2.0 ExtPer output image | 2per image | $0.02per image |
Prices follow the model configuration. Reference images are not billed unless the model says so.
Grok Imagine Image 2.0 is an xAI image generation and editing model for prompt-driven creation, reference-guided revisions, typography, and structured layouts. Use the AIReiter playground or API to generate at 1K or 2K across fourteen aspect options.
Turn a detailed prompt into posters, product scenes, social creative, concept art, and other production-ready visuals. Describe the subject, composition, lighting, style, and any text that should appear.
Attach one to three reference images and describe the required change. The model can use the references to guide subject identity, styling, composition, or prompt-based revisions.
The model is designed for images that combine visual composition with readable labels, headlines, and structured graphic layouts, making it useful for posters and infographics.
Choose 1K or 2K resolution, Low or Medium quality, fourteen aspect options, and one to ten outputs per request through the primary model mode.
Combine natural-language instructions with optional reference images to move from a first concept to a refined result in one workflow.
Use explicit wording, hierarchy, placement, and style instructions when generating posters, menus, labels, or information graphics. Clear prompts help the model preserve the intended visual structure.
Provide up to three references and state exactly what should remain consistent and what should change. Use detailed instructions to make focused revisions while preserving the subject, style, or composition that matters.
Use multiple input images to guide a shared character, product, palette, or visual direction. Reference-image charges are applied once per request for each attached image.
Request several candidates in a single task. The primary mode supports one to ten outputs, while the Ext mode on the same page supports one to twelve outputs.
Use cases that benefit from strong prompt following, structured composition, and reference-guided consistency.
Generate campaign posters, event graphics, menus, and information-rich visuals. Specify the exact copy, layout hierarchy, and visual style in the prompt, then use 2K Medium for final-quality checks.
Create product scenes, ad variations, and user-generated-content-style images. Attach product references when visual identity must stay consistent across new backgrounds or compositions.
Use reference images to maintain a character, wardrobe, palette, or environment while exploring new scenes. Generate several candidates to establish a coherent visual direction for later video work.
These are the controls exposed by the current AIReiter playground and API integration.
Optional for the primary mode. Attach up to three images. Each reference image adds 0.95 credits once per request, regardless of the number of generated outputs.
Choose Auto, 1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, or 9:20 in the primary mode.
The primary mode supports 1K or 2K output with Low or Medium quality. Use Low for faster drafts and Medium when detail matters more.
The primary mode supports one to ten outputs per request. The Ext child mode supports any integer from one to twelve, including values such as 3, rather than only preset batch sizes.
The primary mode uses formula pricing based on resolution, quality, output count, and attached reference images. The Ext child mode uses a flat per-output price.
3.8 credits per output. This is the fastest and lowest-cost tier for drafts and prompt iteration.
5.7 credits per output. Choose 1K Medium for more detail at standard resolution or 2K Low for a larger image with faster quality settings.
7.6 credits per output. This is the highest primary-mode tier for final-quality checks and detail-sensitive work.
Each attached reference image adds 0.95 credits once per request. The surcharge is not multiplied by output_n. Up to three input images are supported.
The output rate is multiplied by output_n. For example, three 1K Low outputs cost 3.8 x 3 credits before any reference-image surcharge.
The Ext child mode costs 2 credits per generated image and supports one to twelve outputs. Its total is 2 x output_n credits.
AIReiter provides one asynchronous task contract for image generation and result retrieval.
Send grok_imagine_image_2_0 as the model value for the full 1K/2K, quality, multi-reference, and one-to-ten-output parameter set.
Send grok_imagine_2_0_ext when you select the Ext child mode. It is text-to-image, supports seven aspect ratios, and accepts one to twelve outputs.
POST /api/openapi/submit accepts model, params, optional attachments, and a unique out_task_id. The response includes the task status and estimated credit cost.
POST /api/openapi/query accepts the same out_task_id. Poll until completed or failed; a completed response includes output URLs and final credits_used.
Compare Grok Imagine Image 2.0 with other generation and editing models available on AIReiter.
A general-purpose image model with 1K to 4K output, custom pixel sizes, and multi-reference editing. Use it when higher resolution or custom dimensions are the priority.
02Nano Banana ProA reference-guided image generation and editing model with broad aspect-ratio support and resolution options up to 4K.
03Seedream 5.0 ProA detailed image generation and editing model with support for larger reference sets and 1K or 2K output.
04Grok Imagine 1.5 VideoUse the Grok video model when the target output is motion rather than a still image, including text-to-video and image-to-video workflows.
| Control | Type | Default | Notes | |
|---|---|---|---|---|
| 01 | prompt | text | — | |
| 02 | image_url | image | — | Upload up to 3 JPG, PNG, or WebP images. |
| 03 | aspect_ratio | string | auto | — |
| 04 | resolution | string | 1K | — |
| 05 | quality | string | medium | — |
| 06 | output_n | integer1 – 10 | 1 | — |
Write a prompt, select the model type, and choose the aspect ratio, resolution, quality, and output count. Add up to three references when using the primary model for an editing workflow.
Create an AIReiter API key and use grok_imagine_image_2_0 as the primary model ID. Keep the key on the server and never expose it in browser code.
Send the task to POST /api/openapi/submit with a unique out_task_id, then poll POST /api/openapi/query with that same ID until the task is completed or failed.
grok_imagine_image_2_0, grok_imagine_2_0_ext
POST to /api/openapi/submit with the model, a unique out_task_id, and parameters.
POST to /api/openapi/query with the same out_task_id until the task completes.
Save outputs in your own storage rather than hot-linking them.
Grok Imagine Image 2.0 is an xAI model for text-to-image generation and prompt-guided image editing. On AIReiter, the primary mode supports optional reference images, 1K and 2K output, Low and Medium quality, and batch generation.
This page generates still images. For motion output, use Grok Imagine 1.5 Video at /video/grok-imagine-1-5.
The current AIReiter API accepts up to three reference images for the primary grok_imagine_image_2_0 mode. The Ext child mode is text-to-image and does not expose reference-image input.
The primary mode offers Auto plus thirteen explicit ratios: 1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, 2:1, 1:2, 19.5:9, 9:19.5, 20:9, and 9:20. Resolution can be 1K or 2K.
Primary-mode output costs 3.8 credits for 1K Low, 5.7 credits for 1K Medium or 2K Low, and 7.6 credits for 2K Medium. Each reference image adds 0.95 credits once per request, and the output rate is multiplied by output_n.
Grok Imagine 2.0 Ext is a child model available from the Model Type selector on this same page. It costs 2 credits per output, supports one to twelve images per request, and uses a smaller text-to-image parameter set.
Both models support general image generation. Grok Imagine Image 2.0 provides Low and Medium quality tiers with formula pricing and up to three references, while GPT-Image 2 offers more resolution and custom-size options with a larger reference-image limit. Choose based on the controls your workflow needs.
Use grok_imagine_image_2_0 for the primary mode or grok_imagine_2_0_ext for Ext. Submit with POST /api/openapi/submit and query the result with POST /api/openapi/query using the same out_task_id.