Alibaba Qwen-Image-2.1 image model opens source
阿里千问开源 Qwen-Image-2.1 图像模型：可生成、编辑透明图像，支持最多 10 张参考图
For users needing a cost-effective, open-source model for transparent image generation and editing, Qwen-Image-2.1 offers a lightweight alternative to proprietary solutions.
Alibaba's Qwen-Image-2.1, an open-source image model, has been released.
It integrates text-to-image generation and image editing.The 7B-parameter model supports generating and editing transparent images.
It can use up to 10 reference images for local edits and is available on GitHub, Hugging Face, and ModelScope.
阿里千问开源 Qwen-Image-2.1 图像模型：可生成、编辑透明图像，支持最多 10 张参考图
IT之家 9 月 20 日消息，阿里千问今日宣布，开源千问图像系列当前兼顾生成效果、推理效率与使用成本的开源图像模型 Qwen-Image-2.1。 Qwen-Image-2.1 将文生图与图像编辑整合在同一模型中，视觉生成部分仅有 7B 参数，并原生支持透明图像的生成与编辑。 官方表示，本次更新的主要特色包括四个方面： 小模强效，极致价比 ：轻量的模型结构与推理优化，在生成质量和计算成本之间取得平衡。 ▲ Qwen-Image-Bench 公开评测对比 原生透明，创改一体 ：根据提示词生成普通图像或透明图像，并支持透明图层编辑与抠图。 ▲ 通过文本直接生成透明图像的示例 全能编辑，多局保全 ：支持最多 10 张参考图 ，增强局部编辑、人像与商品保真，并覆盖多种编辑任务。 ▲ 十张家居参考图生成室内布置 真实质感，字雅人美 ：提升文字排版、人物光影与细节表现，让生成结果更具美感。 ▲ 文字渲染示例 IT之家附相关链接： GitHub： https://github.com/QwenLM/Qwen-Image-2.1 HuggingFace： https://huggingface.co/Qwen/Qwen-Image-2.1 ModelScope： https://www.modelscope.cn/models/Qwen/Qwen-Image-2.1
This report is categorized under AITOP's "Models" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to Chinese AI companies expanding and localizing across Southeast Asia and US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Qwen-Image-2.1, open-source models, image generation, transparent image.
Source attribution: curated from IT之家. Cross-source corroboration: 4 independent sources have covered this development (including 阿里通义 Qwen, Decoder, shao__meng), confirming industry-wide signal strength.
