Qwen has released a new model, Qwen-Image-Edit. Qwen-Image-Edit is the image editing version of Qwen-Image. It is further trained based on the 20B Qwen-Image model, successfully extending Qwen-Image’s unique text rendering capabilities to editing tasks, enabling precise text editing. In addition, Qwen-Image-Edit feeds the input image into both Qwen2.5-VL (for visual semantic control) and the VAE Encoder (for visual appearance control), thus achieving dual semantic and appearance editing capabilities.
According to Qwen Official, Qwen model has key features including:
- Exceptional Text Rendering Capability: Qwen-Image demonstrates outstanding performance in complex text rendering, supporting multi-line layouts, paragraph-level text generation, and fine-grained detail presentation. It achieves high-fidelity output in both English and Chinese.
- Consistent Image Editing Ability: Through an enhanced multi-task training paradigm, Qwen-Image excels at maintaining consistency during the editing process.
- Strong Cross-Benchmark Performance: Evaluations across multiple public benchmarks show that Qwen-Image achieves state-of-the-art (SOTA) results in various generation and editing tasks, establishing it as a powerful foundational model for image generation.
Qwen has conducted comprehensive evaluations of Qwen-Image across multiple public benchmarks, including GenEval, DPG, and OneIG-Bench for general image generation, as well as GEdit, ImgEdit, and GSO for image editing. Qwen-Image achieved state-of-the-art performance across all benchmarks, demonstrating its powerful capabilities in both image generation and editing. Furthermore, results on text rendering benchmarks—LongText-Bench, ChineseWord, and TextCraft—show that Qwen-Image performs exceptionally well in text rendering, particularly in Chinese text rendering, where it significantly outperforms existing state-of-the-art models. This highlights Qwen-Image’s unique position as an advanced image generation model, combining broad general capabilities with outstanding text rendering precision. At the same time, the model handles English and Chinese-English bilingual text with equal proficiency and can seamlessly switch between the two languages.
Qwen's models are open-sourced, and users can download from their Huggingface, Github, ModelScope sites, and get an official workflow from the ComfyUI Template page. However, it may still take some time to download and upload models, as they require a significant amount of space. RunC.AI has made it easier. For those who want to try using Qwen Image and Qwen Image Edit, we have an image in our image community. Users can use it by simply clicking "Deploy" on this image. Remember to read the "README" before connecting to ComfyUI in the image.
Now let's try Qwen Image workflow on RunC.AI!
Firstly, find the image and deploy.
Then, open the workflow in the folder or downloaded and input the prompt you wrote.
Or download the Image Edit workflow from "Readme" and draw the JSON file to the ComfyUI page. Remember to update and restart ComfyUI.
RunC's image community has made creating AI-powered content even more convenient. If you want to generate works of your own without the trouble of downloading the ComfyUI desktop version and other models, just register and try to deploy and run within clicks!
About RunC.AI
Rent smart, run fast. RunC.AI allows users to gain access to a wide selection of scalable, high-performance GPU instances and clusters at competitive prices compared to major cloud providers like Amazon Web Services (AWS), Google Cloud, and Microsoft Azure.
