Z-Image Base

Z-Image Base is an open-source visual generation foundation model for creating and editing AI images online. It is described as a free-to-try image generator, and its page frames it for both individual creators and enterprise teams. The tool uses Alibaba’s Tongyi Lab model and is presented as commercial-ready under the Apache 2.0 license.

Its named functions include text-to-image and image-to-image generation. In text-to-image use, a prompt can be written in natural language, and the tool supports complex scene descriptions, detailed character portraits, and artistic compositions. In image-to-image use, a reference image can be uploaded to generate variations, transform photos into artworks, enhance portraits, and apply new styles. The page also says it supports full CFG classifier-free guidance and negative prompts for controlling visual style and excluding unwanted elements.

Z-Image Base has native bilingual understanding and rendering for Chinese and English prompts and text. It is also described as supporting multiple aspect ratios, including square, portrait, and landscape, with optimization for 1024x1024. The model is said to have 6 billion parameters and to use an S3-DiT architecture that processes text and visual tokens in a unified stream. The page also states that code, weights, and documentation are public, and that generated images are free for commercial use.