Z-Image-Turbo

What is Z-Image-Turbo?

Z-Image-Turbo is an open-source text-to-image model that turns written prompts into photorealistic images fast. It uses a 6B-parameter architecture and an 8-step distilled diffusion process to deliver high-quality results with sub-second latency on enterprise GPUs and good performance on 16GB consumer cards. It renders English and Chinese text accurately, includes a Prompt Enhancer for clearer instruction following, and is released under Apache-2.0 so you can self-host or use a low-cost API. Built for creators, developers, and teams who need fast, reliable image generation.

Key features

• Generates photorealistic images in only eight diffusion steps.
• Sub-second inference on enterprise GPUs and fast consumer performance.
• Accurately renders English and Chinese text directly inside images.
• Runs on 16GB VRAM consumer GPUs for local self-hosted deployment.
• Open-source under Apache-2.0 for commercial and custom use.
• Includes a Prompt Enhancer for better instruction adherence.