Qwen Edit 2511

• 🆕 Alibaba's prompt-driven image editor, built on Qwen-Image 20B architecture.
• 🔧 Performs semantic and appearance edits, modifying or replacing image elements.
• 🧠 Dual encoders: Qwen2.5-VL for semantics, VAE for appearance.
• 🎯 Improves consistency and character preservation over Qwen-Image-Edit-2509.
• 📚 Integrates popular community LoRAs directly into the base model.
• 🌐 Native bilingual (Chinese/English) text editing in images.
• 👁️ Supports multi-image input and multi-subject control.

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 Edit 2511 is Alibaba Tongyi Lab's prompt-driven image editing model, part of the broader Qwen-Image lineage alongside Qwen Image and the newer Qwen Image 2 editing variant. It is built on the 20B-parameter Qwen-Image MMDiT backbone, feeding each input image simultaneously through Qwen2.5-VL for visual semantic control and a VAE encoder for appearance control, enabling both pixel-level appearance edits and higher-level semantic transformations.

According to the official Qwen model card, the 2511 release is an enhanced version over Qwen-Image-Edit-2509, with the team citing notably better consistency, mitigated image drift, improved character consistency, integrated LoRA capabilities, enhanced industrial design generation, and strengthened geometric reasoning. A distinguishing addition over earlier releases is built-in support for selected community-created LoRAs, baked directly into the base model so their effects work without extra tuning.

Functionally, the model supports semantic and appearance editing—modifying or replacing elements within an image—plus Qwen's bilingual text editing for inserting or rewriting Chinese and English text in images. Alibaba's documentation notes the Qwen editing series accepts multiple input images for multi-subject composition.

Within Venice's catalog, Qwen Edit 2511 sits beside an uncensored variant, Qwen Edit Uncensored, and other editing models such as Wan 2.7 Pro Edit. The 2511 model card notes improved character consistency and reduced image drift relative to the prior release.