Z-Image Turbo Image Edit

Z-Image Turbo is a fast 6B model for generating images from text and image inputs using image-to-image technology.

Overview

Z-Image Turbo by Tongyi-MAI is a super-fast 6 billion parameter model designed for image-to-image generation. It accepts both textual prompts and input images to generate high-quality transformed images. Users can control parameters such as image size, inference steps, strength of conditioning, and output format. The model supports prompt expansion and safety-checking features to enhance generation reliability.

Strengths / What it does well

• Generates images from detailed text prompts combined with input images.
• Fast inference with a maximum of 8 inference steps.
• Supports prompt expansion and safety checks to improve outputs.
• Flexible output format options: PNG, JPEG, and WEBP.
• Allows fine control over image-to-image strength for customized transformations.

Best use cases

• Creative image editing and style transfer via text plus input image.
• Generating variations or enhancements on existing images.
• Rapid prototyping of visual concepts combining textual description and images.