Qwen Image Models by Alibaba
4 models2025–2026
Last refreshed 2026-09-21. Next refresh: weekly.
Details
ResearcherAlibaba
LicenseApache 2.0OSI-approved
Commercial useCommercial use: permitted
Models4
Released2025–2026
Capabilities
Vision2 of 4 models
MultimodalAll models
About
Alibaba's Qwen Image family of text-to-image generation models built on Multimodal Diffusion Transformer (MMDiT) architecture. Achieves commercial-grade Chinese and English text rendering. Open-source on HuggingFace (Qwen/Qwen-Image), part of Alibaba's Tongyi/Qwen AI ecosystem.
Decision facts
- Best fit
- image generationimagemultimodal
- Capability starting point
- Qwen-Image-2.1 with multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- Ovis Image
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
4 in view
Qwen-Image-2.1Current
Use when the workload needs image generation, multimodal inputs, and image.
2026-09image generationmultimodal inputsimage
Qwen ImageCurrent
Use when the workload needs image generation, 20B parameters, and multimodal inputs.
2025-08image generation20B parametersmultimodal inputs
Qwen Image EditCurrent
Use when the workload needs image editing and multimodal inputs.
Unknown releaseimage editingmultimodal inputs
Qwen Image MaxCurrent
Use when the workload needs image generation and multimodal inputs.
2025-08image generationmultimodal inputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Qwen-Image-2.1 | Use when the workload needs image generation, multimodal inputs, and image. | 2026-09 | image generationmultimodal inputsimage | Current |
| Qwen Image | Use when the workload needs image generation, 20B parameters, and multimodal inputs. | 2025-08 | image generation20B parametersmultimodal inputs | Current |
| Qwen Image Edit | Use when the workload needs image editing and multimodal inputs. | Unknown release | image editingmultimodal inputs | Current |
| Qwen Image Max | Use when the workload needs image generation and multimodal inputs. | 2025-08 | image generationmultimodal inputs | Current |
Release Timeline
3 release groups2026-09
1 current
Qwen-Image-2.1
Currentimage generationmultimodal inputsimage
2025-08
2 current
Qwen Image
Currentimage generation20B parametersmultimodal inputs
Qwen Image Max
Currentimage generationmultimodal inputs
Unknown release
1 current
Qwen Image Edit
Currentimage editingmultimodal inputs
Specifications(4 models)
| Model | Released | Parameters | Vision | Multimodal |
|---|---|---|---|---|
| Qwen-Image-2.1 | 2026-09 | — | Yes | Yes |
| Qwen Image | 2025-08 | 20B | No | Yes |
| Qwen Image Edit | — | — | Yes | Yes |
| Qwen Image Max | 2025-08 | — | No | Yes |
