Qwen Image 2.1 Pro

• 🏢 Alibaba's Qwen image family, Pro tier, editing/inpainting variant
• 🔧 Instruction-driven image editing via Alibaba Cloud Model Studio's edit API
• 📚 Bilingual English–Chinese text rendering inherited from Qwen-Image-Edit
• 🎯 In-image text can be added, removed or rewritten
• 🧠 Dual encoding: vision-language for semantics, VAE for appearance
• 👁️ Appearance edits keep original font, size and style
• 🆕 Catalog release date: September 2026
• ⚡ A Turbo editing variant also exists in the family

Alibaba

Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research,…

Qwen Image 2.1 Pro (editing variant) is the inpainting and instruction-editing counterpart to Qwen Image 2.1 Pro, Alibaba's Pro-tier text-to-image model. You supply a source image together with a natural-language instruction, and the model rewrites the image accordingly — the Qwen image-edit API on Alibaba Cloud Model Studio is documented for this instruction-based editing workflow.

The family's defining trait is typography. The Qwen-Image-Edit line was built so that text inside an image can be added, deleted or modified while preserving the original font, size and style, with the input routed through both a vision-language encoder for semantic control and a VAE encoder for appearance control — enabling either low-level visual edits or higher-level content changes. Bilingual English and Chinese text handling comes from the same foundation.

Within the catalog, this model sits in the Pro tier of the 2.1 generation, alongside the non-Pro editors Qwen Image 2 and Qwen Image 3, and the separately versioned Qwen Image 3 Pro editor. A faster Turbo editing variant, Qwen Image 2.1 Turbo, is also available for the same instruction-editing workflow. Detailed, independently verified quality comparisons between these tiers are not covered by the primary documentation.