GLM Image
glm-image
Open SourceMultimodal
About
GLM-Image by Zhipu AI (Z.ai). A 16B hybrid model combining a 9B autoregressive language model with a 7B diffusion decoder for text-to-image generation. Pioneered open-source Chinese character and structured text rendering in AI images. The first open-source image model trained entirely on Huawei Ascend hardware. Achieves 91.16% word accuracy on CVTG-2K. Open source, $0.015 per image via Z.ai API.
Capabilities
VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode ExecutionPrompt CachingBatch APIAudioFine-tuning
