GLM-Image is said to bethe first open-source, industrial-grade discrete auto-regressive image generation model, combining a nine-billion-parameter auto-regressive module with a seven-billion-parameter diffusion decoder. Developed by Z.ai (Zhipu AI) in partnership with
Huawei and built on GLM-4-9B-0414 and CogView4, the model reportedly excels in text rendering, semantic accuracy, and complex image-to-image tasks.