Qwen Image
Frontierby Alibaba / Qwen
State-of-the-art text-to-image model from Qwen team. 20.4B DiT transformer with Qwen2.5-VL (8.3B) text encoder. Excels at photorealism, Chinese/English text rendering, and complex compositions. Apache 2.0 licensed.
VRAM requirements, GPU fit, and setup notes for Qwen Image, including 8GB/12GB fit guidance where relevant. Recommended runtimes: ComfyUI and Diffusers support. Best download size: ~26.7 GB at FP8.
- 20.4B DiT — largest open text-to-image model
- Qwen2.5-VL (8.3B) as text encoder for superior text understanding
- Excellent Chinese and English text rendering
- Apache 2.0 — fully open for commercial use
Your hardware
Detecting...
Image Quality Benchmarks
Measured quality metrics for Qwen Image outputs.
How often humans prefer this model's output (0-100%)
Visual quality and composition rating (5-9 scale)
VRAM Requirements by Resolution and Precision
Compare which GPUs can run Qwen Image at different precisions. FP8 uses less memory than FP16 when available, and the grade shows how comfortably each GPU handles the workload.
FP16 (full precision)
| Resolution | VRAM Required | RTX 4090 24GB | RTX 3060 12GB | RTX 4060 8GB | MacBook Pro M4 Pro 24GB |
|---|---|---|---|---|---|
| 512×512 | 61.2 GB | F | F | F | F |
| 768×768 | 61.4 GB | F | F | F | F |
| 1024×1024 | 61.8 GB | F | F | F | F |
FP8 (~40% less VRAM)
| Resolution | VRAM Required | RTX 4090 24GB | RTX 3060 12GB | RTX 4060 8GB | MacBook Pro M4 Pro 24GB |
|---|---|---|---|---|---|
| 512×512 | 23.5 GB | B● | F● | F● | F● |
| 768×768 | 23.7 GB | B● | F● | F● | F● |
| 1024×1024 | 24.0 GB | B● | F● | F● | F● |
Optimization Tips
Turbo / LCM distillation
Use distilled scheduler at 4-8 steps for faster iteration
Run with Python
from diffusers import DiffusionPipeline
import torch
pipe = DiffusionPipeline.from_pretrained(
"Qwen/Qwen-Image",
torch_dtype=torch.float16
)
pipe.to("cuda")
image = pipe(
prompt="your prompt here",
num_inference_steps=30,
guidance_scale=5.0,
height=1024,
width=1024,
).images[0]
image.save("output.png")Get started
Setup instructions for running Qwen Image locally
1. Download the model
Get the checkpoint from HuggingFace
2. Place in:
ComfyUI/models/checkpoints/3. Launch ComfyUI
python main.pyMemory Breakdown
VRAM allocation at 1024×1024 on RTX 4090 24GB (24 GB)
Estimated Generation Time
Time per image at 1024×1024, 28 steps, FP16.
Sample Outputs
Available Formats, Downloads & Setup
Download Qwen Image in the precision that matches your GPU. Lower precision usually means less VRAM pressure, while higher precision keeps more quality.
| Format | Präzision | Größe | Anbieter | |
|---|---|---|---|---|
| Offizielle Gewichte | ||||
| safetensors | BF16 | 53.4 GB | official | Herunterladen |
| Community-Konvertierungen | ||||
| safetensorsCommunity | FP8 | 26.7 GB | community | Herunterladen |
LoRA Ecosystem
growingRapidly growing LoRA ecosystem as one of the newest frontier image models.
Related Workflows
You might also like
Frequently asked questions
FAQ — Qwen Image VRAM, Runtimes & Fit
How much VRAM does Qwen Image need?
Qwen Image (20.4B parameters) requires approximately 61.8 GB of VRAM at FP16 precision for standard 1024×1024 image generation. If you want a lighter setup, lower precisions like FP8 can reduce memory use when available.
Can I run Qwen Image on an 8GB GPU?
Qwen Image usually needs more than 8GB for comfortable local use. Check the VRAM table above for the exact resolution and precision trade-off.
Does Qwen Image work in ComfyUI and Diffusers?
Qwen Image is marked for ComfyUI and Diffusers support in our catalog, so those are the runtimes we recommend first for local setup. If your workflow uses another front end, check the model's available formats and workflow notes above before downloading.
Can I run Qwen Image on RTX 4090?
Qwen Image is too large for the RTX 4090's 24 GB VRAM at FP16. Consider using FP8 precision or a GPU with more VRAM.
Does Qwen Image support ControlNet?
There are currently no known ControlNet adapters for Qwen Image. Check Hugging Face and Civitai for community-contributed adapters.
Does Qwen Image have LoRA support?
Rapidly growing LoRA ecosystem as one of the newest frontier image models. The LoRA ecosystem for Qwen Image is rated as "growing". Each LoRA adds roughly 0.4 GB of extra VRAM.
How fast is Qwen Image?
On a reference GPU (RTX 4090 24GB), Qwen Image generates a 1024×1024 image in approximately ~6.7s at FP16 with 28 inference steps. Faster GPUs with higher memory bandwidth will produce images more quickly.
About Qwen Image
See also