88Z AI, 21 September — Alibaba's Tongyi team has released the open-source image model Qwen-Image-2.1. With only 7 billion parameters in its visual generation component, it balances output quality with inference costs. The model integrates text-to-image and image editing capabilities, supporting transparent image generation and local edits, while significantly improving fidelity for portraits and products.
At a glance
- Visual generation uses only 7B parameters, balancing quality and cost
- Supports transparent image generation and various local editing methods
The update
Alibaba's Tongyi team has officially open-sourced the Qwen-Image-2.1 image model. The visual generation component of this model contains only 7 billion parameters, aiming to achieve an effective balance between high-quality generation and low inference costs. Qwen-Image-2.1 deeply integrates text-to-image and image editing functions, natively supporting transparent image generation, transparent layer editing, and matting operations. Regarding input and editing, the model supports up to 10 reference images and provides various local editing methods such as selection, brushing, and masking. These improvements have significantly enhanced fidelity when processing portrait and product images, allowing for more precise restoration of details to meet user needs.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 千问大模型Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.