Alibaba: qwen-image-plus-2026-01-09 API

Qwen-Image-Edit-Max is the most powerful flagship image editing model in the Tongyi Wanxiang (Qwen-Image) series. Built on a 20B-parameter multimodal diffusion architecture (MMDiT), it utilizes an innovative framework that feeds input images into both Qwen2.5-VL (for visual semantic control) and a VAE Encoder (for visual appearance control). Compared to the Plus version, the Max model significantly enhances the understanding of physical laws (e.g., auto-generating matching shadows and reflections), geometric reasoning for industrial design, and character consistency. It supports multi-image fusion (up to 3 images), enabling precise bilingual text modification, object addition/removal, pose changes, and style transfer, while greatly mitigating image drift issues.

• Context window: 2,048 tokens
• Input: text
• Output: image
• Released: 2026-01-15
• Knowledge cutoff: 2025-11-30