The Year in AI Papers
The year/Qwen

Paper 2605.10730

Qwen-Image-2.0 Technical Report

Bing Zhao, Chenfei Wu, Deqing Li, Hao Meng, Jiahao Li, Jie Zhang, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kuan Cao, Kun Yan, Liang Peng, et al.

Published
May 2026
Research lab
Qwen
Citations
8
GitHub
Not linked

01 In brief

Summary

Qwen-Image-2.0 is an image generation foundation model that unifies text-to-image (T2I) generation and instruction-based image editing in a single framework.

It addresses challenges in ultra-long text rendering (up to 1K tokens), multilingual typography, high-resolution photorealism (native 2K), complex instruction following, and inference efficiency.

The architecture couples a frozen Qwen3-VL encoder with a Multimodal Diffusion Transformer (MMDiT) and a high-compression VAE (16x downsampling, f16c64).

Training uses a multi-stage pipeline (pre-training, continual pre-training, SFT, RLHF) with a resolution curriculum from 256p to 2048p, and a data flywheel for iterative improvement.

A Prompt Enhancer (PE) module rewrites user prompts for better generation.

RLHF uses task-specific reward models (aesthetic, alignment, portrait, instruction-following, visual consistency) optimized via GRPO.

Few-step distillation (DMD) produces a 4-NFE student model.

On LMArena, Qwen-Image-2.0 achieves an ELO of 1168, ranking #9 globally and #1 among Chinese models, outperforming Nano Banana.

Qualitative comparisons show superior text rendering, portrait generation, and identity preservation in editing tasks compared to GPT-Image-2, NanoBanana Pro, Qwen-Image-2512, Wan2.7 Pro, and Seedream 5.0 Lite.

The model demonstrates significant improvements over previous Qwen-Image series in visual quality, editing capability, and usability.

The report includes detailed data curation, architecture, training, and evaluation sections, with contributions from a large team of researchers.

02 From the paper

Abstract

We present Qwen-Image-2.0, an omni-capable image generation foundation model that unifies high-fidelity generation and precise image editing within a single framework. Despite recent progress, existing models still struggle with ultra-long text rendering, multilingual typography, high-resolution photorealism, robust instruction following, and efficient deployment, especially in text-rich and compositionally complex scenarios. Qwen-Image-2.0 addresses these challenges by coupling Qwen3-VL as the condition encoder with a Multimodal Diffusion Transformer for joint condition-target modeling, supported by large-scale data curation and a customized multi-stage training pipeline. This enables strong multimodal understanding while preserving flexible generation and editing capabilities. The model supports instructions of up to 1K tokens for generating text-rich content such as slides, posters, infographics, and comics, while significantly improving multilingual text fidelity and typography. It also enhances photorealistic generation with richer details, more realistic textures, and coherent lighting, and follows complex prompts more reliably across diverse styles. Extensive human evaluations show that Qwen-Image-2.0 substantially outperforms previous Qwen-Image models in both generation and editing, marking a step toward more general, reliable, and practical image generation foundation models.