# GLM Image AI Generator and Editor

> A hybrid autoregressive image model for dense knowledge, accurate text and multi-image editing.

Canonical URL: https://allimageai.com/models/glm-image/

Website: All Image AI

Language: English

GLM Image combines an autoregressive visual-token generator with a diffusion decoder. It is built for information-dense graphics, accurate text and knowledge-heavy prompts, while the matching edit operation accepts as many as four source images.

## Model details

- Developer: Z.ai
- Status on All Image AI: Active
- Last verified: 2026-08-25
- Maximum output: 2K
- Modes: Text to Image, Image to Image, Edit Image, Multi-reference, Typography
- Reference images: up to 4
- Resolutions: 1K, 2K
- Aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3

## Best for

- Infographics and knowledge-rich visuals
- Accurate in-image text
- Multi-subject editing
- Identity and object consistency

## How to use

1. Put exact words in quotation marks.
2. Upload up to four ordered references when editing.
3. Iterate at 1K.
4. Use 2K for the selected information-dense final.

## Settings explained

### Guidance

The documented default is 1.5; higher values follow literal wording more strongly.

### Steps

Start at 30 for drafts and use more only for a difficult final composition.

## Prompt examples

### Recipe infographic

> Clean magazine infographic titled “RASPBERRY MOUSSE”, ingredient list on the left, four photographed steps on the right, cream and raspberry palette, exact readable labels

### Multi-image edit

> Keep the person from image 1, dress them in the coat from image 2, place them in the room from image 3, and preserve the product from image 4

## Limitations

- The full model is slower than distilled generators
- 2K costs four times the 1K pixel tier
- Prompt expansion is disabled so pricing and wording stay predictable

## Pricing and credits

Credits for one image at default quality and mode. Resolution, quality, reference inputs, edit mode and output count can change the total. Check the Generate button before submitting.

| Resolution | Credits / image |
| --- | --- |
| 1K | 4 |
| 2K | 15 |

Failed generations cost 0 credits. Catalog availability does not guarantee runtime availability; the workspace reports whether the selected model can run.

[Credit plans](https://allimageai.com/pricing.md)

## Test results

Testing in progress. No verified benchmark scores have been published for this model.

## Commercial use and image handling

GLM Image is released under Apache 2.0; generated use remains subject to applicable law and service terms.

Uploads are processed by the configured inference service. You can delete a generation from your history.

## Frequently asked questions

### Why is GLM Image useful for infographics?

Its architecture is designed around dense knowledge and glyph fidelity rather than treating text as ordinary texture.

### Does multi-reference use the same model?

Yes. The integration selects GLM Image’s documented image-to-image companion and sends up to four images.

## Release history and sources

January 14, 2026 — Model announced · open weights

This date applies to the event shown. Later API access, weights and availability on All Image AI may have different dates.

- [z.ai](https://z.ai/blog/glm-image)

[Event evidence on the timeline](https://allimageai.com/timeline/#event-glm-image)