#4242nd of 42 in Open modelsFor image, video and design, image and design
qwen-image
20B MMDiT image generation and editing model with strong text rendering
Live demo ↗
(opens in a new tab)Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 8.4k
- License
- Apache-2.0
- Last commit
- Feb 2026
- Language
- Python
7,582 on 1 Apr 20268,422 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Qwen-Image is a 20B MMDiT image foundation model from the Qwen team, covering text-to-image, single and multi-image editing, and layered output. The repo holds diffusers quick-start code, prompt-enhancement helpers that call DashScope, and a Gradio demo server that spreads work across multiple GPUs. Weights are on Hugging Face and ModelScope.
Who it is for: Engineers running or building on open-weight image generation and editing models
Strengths
- Text-to-image and editing variants (2512, Edit-2511, Layered) published as open weights
- Strong text rendering, with Chinese called out specifically
- Runs through diffusers pipelines; ComfyUI, vLLM-Omni and SGLang-Diffusion support noted
- Gradio server with multi-GPU workers and a task queue, set by env vars
Weaknesses
- VRAM and RAM requirements for a 20B model are not stated in the README
- Needs diffusers installed from git main and transformers>=4.51.3
- Prompt enhancement needs a DashScope API key, an external paid service
- Qwen-Image-2.0 is announced for Qwen Chat; README does not link open weights
What it needs
- GPU optional
- Needs diffusers, transformers>=4.51.3, PyTorch, DashScope API (optional, prompt enhancement)
- Models: Qwen-Image, Qwen-Image-2512, Qwen-Image-Edit, Qwen-Image-Edit-2509, Qwen-Image-Edit-2511
Also in Open models
See all 42| Rank | Project | Score |
|---|---|---|
| 1 | llmfitRust tool that checks which LLMs fit your RAM, CPU and GPU | 87 out of 100 |
| 2 | layaSingle-pass multilingual classifier for typed choice, score and yes/no decisions | 86 out of 100 |
| 3 | timesfmPretrained time-series foundation model for zero-shot forecasting from Google Research | 76 out of 100 |
| #4 | speechPyTorch toolkit for training and deploying ASR, TTS and speech LLM models | 73 out of 100 |
| #5 | yueLyrics-to-song model that writes an editable melody and chord score first | 69 out of 100 |