Z Image Edit

Z-Image Model Capabilities

⚡ Lightning Turbo Inference

Z-Image-Turbo delivers sub-second generation with only 8 NFEs, running efficiently on consumer hardware with 16G VRAM

🏆 Photorealistic Quality

Matches or exceeds leading competitors in image realism, detail, and visual fidelity

🌍 Perfect Bilingual Text

Flawless Chinese and English typography rendering, even in complex layouts and small fonts

🧠 Deep World Knowledge

Built-in reasoning capabilities for accurate celebrity, landmark, and cultural generation

✏️ Advanced Image Editing

Z-Image-Edit provides precise instruction-following edits with natural language prompts

🔧 Community Fine-tuning

Z-Image-Base enables full potential for community-driven fine-tuning and custom development

🚀 Enterprise Performance

Optimized for H800 GPUs with Flash Attention for maximum speed and efficiency

💎 Open Source Power

Apache 2.0 licensed models with complete access to checkpoints and training code

Z-Image Model Variants

🚀 Z-Image-Turbo – A distilled version that matches or exceeds leading competitors with only 8 NFEs. Offers ⚡️sub-second inference latency⚡️ on enterprise-grade H800 GPUs and fits comfortably within 16G VRAM consumer devices. Excels in photorealistic image generation, bilingual text rendering, and robust instruction adherence.

🧱 Z-Image-Base – The non-distilled foundation model. By releasing this checkpoint, we aim to unlock the full potential for community-driven fine-tuning and custom development.

✍️ Z-Image-Edit – A variant fine-tuned specifically for image editing tasks. Supports creative image-to-image generation with impressive instruction-following capabilities, allowing for precise edits based on natural language prompts.