Z.aiGLM Image
GLM-Image
A hybrid image generator with separate semantic and diffusion components.
Practical use
Practical guidance. For text-heavy artwork, review spelling and the relationship between labels and objects. A well-rendered label can still identify the wrong part. Check the hosted endpoint's accepted inputs before assuming it exposes every local editing capability.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Downloadable |
| Model identifiers | glm-imagezai-org/GLM-Image |
| Architecture | Nine-billion-parameter autoregressive generator initialized from GLM-4-9B-0414; seven-billion-parameter diffusion decoder with a Glyph Encoder.GLM-Image official documentation |
| Weights | Support text-to-image and image-to-image tasks under MIT.GLM-Image official documentation |
| Hosted access | Current API catalogue lists GLM-Image.Current Z.ai API catalogue |