Accurate Bilingual Text Rendering in AI Images

Z-Image-Turbo utilizes a Scalable Single-Stream DiT (S3-DiT) architecture, trained to accurately render text in both English and Chinese.

You can render any text in both English and Chinese, from simple phrases to more complex sentences and brand names.

Yes, Z-Image-Turbo generates high-quality images suitable for various commercial applications, including advertising, marketing, and product visualization.

Z-Image-Turbo comfortably runs on consumer devices with 16GB VRAM and achieves sub-second inference on enterprise H800 GPUs.