Grok Imagine Image 2.0 (grok-imagine-image-2.0) is the main image model on the xAI API. Since November 2, 2026, requests to grok-imagine-image-quality and grok-imagine-image-pro are served by Image 2.0 at low quality, for $0.04 per 1K image. Switch the model name explicitly to control quality and cost.
Last updated: November 8, 2026
Table of contents
- What changed on November 2
- Pricing compared with earlier models
- Choosing low, medium or auto
- How to migrate step by step
- Features worth using
- Limits and availability
- Frequently asked questions
- Sources
What changed on November 2
The xAI API release notes record that grok-imagine-image-quality and grok-imagine-image-pro are now served by grok-imagine-image-2.0-low, meaning grok-imagine-image-2.0 with quality set to low. xAI gave 60 days of notice, starting September 2, 2026.
The migration guide explains how the redirect behaves:
- It applies to both /v1/images/generations and /v1/images/edits.
- Request and response shapes are unchanged, so every parameter you sent before is still accepted.
- usage.cost_in_usd_ticks on each response reflects the lower Image 2.0 price, which is the quickest way to confirm the switch.
- gRPC responses from the xAI SDK report the model as grok-imagine-image-2.0-low.
- grok-imagine-image (1.0) is not affected.
Because the slug keeps working, nothing breaks. The catch is that you no longer choose the quality level. Every redirected request is served at low, even editing requests where the default for Image 2.0 would otherwise be medium.
Pricing compared with earlier models
The figures below come from the migration guide and the xAI pricing page. Prices are per output image.
| Model or setting | Status | Published price |
|---|---|---|
| grok-imagine-image-2.0, quality low, 1K | Current | $0.04 per image |
| grok-imagine-image-2.0, quality low, 2K | Current | $0.06 per image |
| Image edits on Image 2.0 (low) | Current | Output price plus $0.01 per input image |
| grok-imagine-image-quality | Retired November 2, 2026 | $0.01 per image more than Image 2.0 low at every resolution |
| grok-imagine-image-pro | Retired; redirects to Image 2.0 low | Billed at the Image 2.0 low rate |
| grok-imagine-image (1.0) | Current, not affected | $0.02 per image |
The pricing page lists a single $0.04 base rate for grok-imagine-image-2.0 and does not print a separate medium price. Images are billed at the quality they are served at, so check cost_in_usd_ticks on a few medium requests before you scale them up. Image generation in the Batch API is billed at standard rates, without the batch discount that applies to text models. For token pricing on Grok text models, see our Grok API pricing answer.
Choosing low, medium or auto
The image generation docs describe an optional quality parameter that only grok-imagine-image-2.0 supports:
- low: the redirect setting and the cheapest option. xAI says it ranks second on the Arena Image Edit leaderboard and third on Text-to-Image.
- medium: more compute per image for finer detail, for work where quality matters more than cost.
- auto: the default when quality is omitted. It currently serves low for generation and medium for editing.
If your budget assumed the retired model, pinning low keeps costs predictable. If you edit product photos or portraits, test medium on a sample set first.
How to migrate step by step
- Search your code and configuration for grok-imagine-image-quality and grok-imagine-image-pro.
- Change the model field to grok-imagine-image-2.0. xAI describes this as a one-line change.
- Add quality explicitly. In the OpenAI Python SDK, pass it through extra_body, for example extra_body={“quality”: “low”}. The xAI SDK accepts quality=”low” on client.image.sample.
- Apply the same change to editing calls on /v1/images/edits. The OpenAI SDK images.edit() method is not supported there because xAI expects JSON, so use the xAI SDK, the Vercel AI SDK or direct HTTP requests.
- Run a sample of real prompts, compare output, and log cost_in_usd_ticks to confirm the bill matches the quality you chose.
Features worth using
Image 2.0 accepts more than the model it replaces. Since the August update listed in the release notes, it takes up to five source images per edit (previously three) and two extra aspect ratios, 21:9 for cinematic widescreen and 5:2 for wide banners. Other options documented for generation:
- Batches: n from 1 to 10 images per request. The xAI Python SDK uses sample_batch(n=…).
- Resolution: 1k by default, or 2k.
- Aspect ratio: auto by default, or a fixed ratio such as 1:1, 16:9, 9:16 or the 19.5:9 and 20:9 phone formats.
- Output format: temporary URLs by default, or b64_json for direct embedding.
- Deferred generation: set deferred to true, poll GET /v1/images/{request_id}, or register webhooks such as image.generation.completed. Results are kept for 24 hours.
Generated URLs are temporary, so download outputs promptly or use storage options to save them to the Files API.
Limits and availability
The grok-imagine-image-2.0 model page lists text and image input with image output, Batch API support, a rate limit of 6 requests per second, and availability in the us-east-1 and us-west-2 regions. Priority Processing does not apply to image generation, according to the pricing page.
Other recent Grok API changes affect text and voice models rather than images. See our guides to Grok API output watermarking, custom moderation controls and Grok Voice 3.0.
Frequently asked questions
Is grok-imagine-image-quality still available?
The slug still resolves, but the model behind it is retired. Since November 2, 2026, requests to grok-imagine-image-quality are served by grok-imagine-image-2.0 with quality set to low, so existing code keeps working without changes.
How much does Grok Imagine Image 2.0 cost?
The migration guide lists $0.04 per 1K image and $0.06 per 2K image at low quality, plus $0.01 per input image on edits. The pricing page lists grok-imagine-image-2.0 at $0.04 per image and the older grok-imagine-image at $0.02 per image.
What happened to grok-imagine-image-pro?
grok-imagine-image-pro was already redirected to grok-imagine-image-quality during the May 15, 2026 retirement. It now follows that redirect to grok-imagine-image-2.0 at low quality.
Should I use low, medium or auto quality?
Low is the cheapest setting and what the redirect uses. Medium spends more compute per image for finer detail. Auto, the default, currently serves low for generation and medium for editing, and you are billed at the quality actually served.
Is grok-imagine-image 1.0 affected?
No. xAI says grok-imagine-image (1.0) is not affected by the November 2 change. It remains listed on the pricing page at $0.02 per image.
Sources
- xAI API release notes (November 2, 2026 entry)
- grok-imagine-image-quality retirement migration guide
- xAI image generation documentation
- xAI API pricing
- grok-imagine-image-2.0 model page
Image: Grok logo via Wikimedia Commons, public domain. Grok is a trademark of SpaceXSI (xAI). FreeAnswers is not affiliated with or endorsed by xAI.
