Post-Cutoff
Search For AI agents

Alibaba (Qwen)Qwen-Image familyCurrentImage generationReleased 20 Sep 2026

Qwen-Image-2.1

Quick start (from the README):

Model idQwen/Qwen-Image-2.1

PricePrice not published

Checked10 October 2026

Where it stands

6newer image generation models since its release

The 5 newest are listed. All image generation models

Before it from Alibaba
Qwen-Image-3.0 (Pro)21 Jul 2026
Weights
Weights are public. Licence: Qwen Research License. See the licence for commercial terms.

How to call it

5 ways to use Qwen-Image-2.1, with the model id each provider expects.

Providers, model ids, endpoints and docs for Qwen-Image-2.1
ProviderModel idEndpointDocs
Hugging Face Qwen/Qwen-Image-2.1 huggingface.co
ModelScope modelscope.cn
Hugging Face (Turbo, 8-step) Qwen/Qwen-Image-2.1-Turbo huggingface.co
GitHub github.com
Hugging Face Space (demo) huggingface.co

Pricing

Price not published.

No price source linked. Checked 10 October 2026.

Specs

Input
Text and images
Output
Images
Context window
Not published
Max output
Not published
Open weights
Yes
Licence
Qwen Research License
Released
20 September 2026
Status
Current
Type
Image generation

What stands out

What its maker marketed at launch, and what people found later, each with its source.

  • Native RGBA generation and editing

    Generates transparent images, edits transparent layers and extracts subjects from photos in one model.

    Source: github.com

  • Multi-reference editing

    Up to 10 reference images; local edits via circles, painted annotations or masks, with identity preservation.

    Source: github.com

Notes

7B visual generator (32 single-stream DiT layers). Diffusers pipeline QwenImage21Pipeline (install diffusers from git); day-0 ComfyUI, vLLM-Omni, SGLang support. Qwen Research License, check terms before commercial use. Qwen-Image-2.1-Turbo (HF, 2026-10-09): accelerated checkpoint, 8 denoising steps, CFG=1, same 7B architecture and pipeline, same qwen-research license; schedule saved in the checkpoint (needs Diffusers with pipeline-configured sigmas, PR #14950).

import torch
from diffusers import QwenImage21Pipeline
pipe = QwenImage21Pipeline.from_pretrained("Qwen/Qwen-Image-2.1", torch_dtype=torch.bfloat16).to("cuda")
image = pipe(prompt="A neon shop sign that reads \"QWEN IMAGE 2.1\"", num_inference_steps=40).images[0]