What GPU do I need to run Qwen/Qwen-Image?
A 20.4B-parameter text-to-image model. 20.4B parameters, published in BF16. View on Hugging Face
Qwen-Image is published by Qwen on Hugging Face, with 289,743 downloads and 2,601 likes to date. It's a unlisted-architecture model built for text-to-image, published natively in BF16.
What Qwen-Image is
Qwen-Image is a 20.4B-parameter text-to-image model published by Alibaba (Qwen) on Hugging Face, released under Apache 2.0.
License note: fully permissive, no gating and no usage restrictions. Facts in this section are sourced from Qwen-Image's Hugging Face model card, not benchmarked by Aquanode.
What it's used for
- Text-to-image generation
- Creative asset generation
- ComfyUI workflows
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for activation memory and allocator fragmentation. Diffusion and video models carry no KV-cache. The real driver of extra memory is output resolution and frame count, which this flat overhead does not model. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 38.1 GB | 45.7 GB | RTX A6000 | 1 | $0.363/hr |
| cheaper alt. | RTX A5000 | 2 | $0.352/hr | ||
| FP8 (quantized) | 19.0 GB | 22.8 GB | RTX 4090 | 1 | $0.374/hr |
| INT4 (quantized) | 9.5 GB | 11.4 GB | RTX 3060 | 1 | $0.110/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run Qwen-Image at its published (BF16) precision: 1× RTX A6000, at $0.363/hr per GPU ($0.363/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Qwen-Image: common questions
Can Qwen-Image run on a single GPU?
Yes, but not on a desktop card. At BF16 it needs 45.7 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest live single-card fit is a 48.0 GB RTX A6000 at $0.363/hr.
What is the least VRAM Qwen-Image can run in?
11.4 GB, at INT4 (quantized), which fits a 12 GB card, against 45.7 GB at BF16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Does quantizing Qwen-Image lower the GPU bill?
Yes. At BF16 the cheapest live fit is one RTX A6000 at $0.363/hr. At INT4 (quantized) it drops to one RTX 3060 at $0.110/hr, provided a quantized checkpoint exists for it.
How to run Qwen-Image
Run Qwen-Image with Diffusers (Python)
Generic example using Hugging Face's diffusers library, not from the model's own docs.
from diffusers import DiffusionPipeline
import torch
pipe = DiffusionPipeline.from_pretrained("Qwen/Qwen-Image", torch_dtype=torch.bfloat16)
pipe.to("cuda")
image = pipe("a description of the scene").images[0]
image.save("output.png")Run Qwen-Image with ComfyUI
Aquanode's ComfyUI template comes with ComfyUI preinstalled. Download Qwen/Qwen-Image's checkpoint into the models folder and load it in a workflow; this is a real Aquanode template, but loading this specific checkpoint is a manual step, not a one-click deploy.
Deploy Qwen-Image on Aquanode
Aquanode has no one-click deploy template for Qwen-Image; it comes with ComfyUI preinstalled, so you only need to load the checkpoint, not install anything. Aquanode sells GPU pods billed per second, not a hosted inference API.
- Launch the ComfyUI template sized to the requirement above (1× RTX A6000 or larger).
- Open a terminal on the pod, or save one of the commands above as a startup script so it runs automatically the first time the pod boots.
- Run the command and connect to the resulting endpoint.
Launch the ComfyUI templateWeight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More Qwen-Image models
- Qwen-Image-Edit-2511 (20.4B, BF16)
- Qwen-Image-2512 (20.4B, BF16)
- Qwen-Image-Bench (27.4B, BF16)
- Qwen-Image-2.1-viggle-turbo (7.1B, BF16)
- Qwen-Image-2.1 (7.1B, BF16)
Alternatives at this size
Other models for text-to-image within about a third of Qwen-Image's 20.4B parameters, from other model lines.
- HiDream-I1-Fast (17.1B, BF16)
More on Qwen-Image
Fits on a 48 GB GPU at BF16: every model that fits in 48 GB.
Fits on a 24 GB GPU at FP8: every model that fits in 24 GB.
Fits on a 12 GB GPU at INT4: every model that fits in 12 GB.
Best image generation models: how Qwen-Image ranks against the rest.
Compare Qwen-Image with stable-diffusion-xl-base-1.0, sdxl-turbo, FLUX.1-dev, FLUX.1-schnell, stable-diffusion-v1-4, sd-turbo, stable-diffusion-3.5-medium, stable-diffusion-3.5-large.
Related reading: RTX A6000 pricing and specs, Fast diffusion inference on GPU VMs, and What AI inference is.
