Jfbjuj/z-image-turbo-playground

Z-Image Turbo Playground

Create high-fidelity images from natural language with Tongyi-MAI/Z-Image-Turbo.

This Space contains a ZeroGPU-ready Gradio app using the official Diffusers pipeline and the model's recommended Turbo settings: guidance_scale=0.0 and 9 inference steps (8 DiT forwards).

The Space is staged as a free Static Space while a community GPU grant is reviewed. After the grant is approved, change sdk: static to sdk: gradio, change app_file: index.html to app_file: app.py, and add suggested_hardware: zero-a10g.

Notes

• —The first generation on a cold worker may take longer while the model is loaded.
• —Z-Image-Turbo is optimized for photorealism, bilingual text rendering, and strong prompt adherence.
• —This demo is intended for creative exploration. Please follow the model's license.