The core features of Z-Image are **few parameters (6B), fast speed, and high quality**.

Its distillation optimized version, Z-Image-Turbo, can generate images with photo-realistic details in just **8 inference steps**, excelling in skin texture, hair details, and lighting rendering. It is especially good at **high-precision text rendering in both Chinese and English** and can accurately understand complex natural language instructions.

Currently, it does not support image generation on the Weiui end; please use the workflow for image generation, as follows:

Z-image-Model Image Generation Workflow