<h1>QuantFunc A4W4 INT4 — quantized from Qwen-Image-2.1</h1>
<p><strong>True 4-bit inference — A4W4 (4-bit activations × 4-bit weights).</strong></p>
<p>QuantFunc's core quantized matrix multiplications use <strong>INT4 activations and INT4 weights</strong> on its INT4 inference backend. A4W4 describes the compute precision used during inference, alongside the reduced weight size and memory bandwidth demand.</p>
<p>QuantFunc INT4 compresses <a href="https://huggingface.co/Qwen/Qwen-Image-2.1" rel="ugc">Qwen-Image-2.1</a> to about a quarter of its 16-bit size. In our visual comparisons, composition, text rendering, detail and style stay close to the 16-bit baseline.</p>
<h2>Showcase</h2>
<p>All images below were generated by Qwen-Image-2.1-QuantFunc-4bit.</p>



Floral ad design
Portrait photography




<img alt="Floral ad design" src="https://huggingface.co/QuantFunc/Qwen-Image-2.1-4bit/resolve/main/assets/qwen21_show.webp" />
<img alt="Portrait" src="https://huggingface.co/QuantFunc/Qwen-Image-2.1-4bit/resolve/main/assets/qwen21_show2.webp" />


<strong>Chinese food poster</strong>
<strong>Whimsical everyday scene</strong>


<img alt="Chinese food poster" src="https://huggingface.co/QuantFunc/Qwen-Image-2.1-4bit/resolve/main/assets/qwen21_show3.webp" />
<img alt="Whimsical everyday scene" src="https://huggingface.co/QuantFunc/Qwen-Image-2.1-4bit/resolve/main/assets/qwen21_show4.webp" />



<h2>Swap one loader, keep your workflow</h2>
<ol>
<li>Install or update <a href="https://github.com/QuantFunc/ComfyUI-QuantFunc" rel="ugc">ComfyUI-QuantFunc</a>.</li>
<li>Download the r128 or r32 weights.</li>
<li>Swap your model loader for the QuantFunc loader and pick the matching weight file.</li>
</ol>
<p>Every other node, connection and generation parameter stays as-is.</p>
<h2>RTX 20-series through GB300, one build covers it all</h2>
<p>Runs on every NVIDIA <strong>SM75+</strong> GPU: RTX 20/30/40/50-series, A100, H100, H200, B100, B200, GB300.</p>
<p>4-bit weights significantly cut the weight-bandwidth cost of loading and inference, making Qwen-Image-2.1 much easier to run on consumer GPUs. Actual VRAM needs depend on resolution, batch size and the rest of your workflow.</p>
<h2>Choose a model</h2>



Variant
File
Size
Use case




<strong>r128</strong>
<code>qwen_image_21_base_quantfunc_int4_r128.safetensors</code>
4.66 GB (4.34 GiB)
Quality-first, recommended


<strong>r32</strong>
<code>qwen_image_21_base_quantfunc_int4_r32.safetensors</code>
4.33 GB (4.03 GiB)
Size/VRAM-first



<h2>Loading</h2>
<p>Weights use QuantFunc's own safetensors format — load with <strong>ComfyUI-QuantFunc</strong> or the <strong>QuantFunc inference engine</strong>, not as a drop-in Diffusers checkpoint.</p>
<p>If the model isn't recognized or fails to load, update ComfyUI-QuantFunc to the latest version and restart ComfyUI.</p>
<h2>Source &amp; license</h2>
<p>This repository distributes derived quantized weights produced from <a href="https://huggingface.co/Qwen/Qwen-Image-2.1" rel="ugc">Qwen/Qwen-Image-2.1</a>, released by the <a href="https://huggingface.co/Qwen" rel="ugc">Qwen</a> team.</p>
<p>Qwen-Image-2.1 and its derivative weights follow the <strong>Qwen Research License</strong> (<a href="https://huggingface.co/Qwen/Qwen-Image-2.1/blob/main/LICENSE" rel="ugc">license text</a>). QuantFunc's quantization tooling and implementation follow their own respective licenses. Please read and comply with the original model's license terms before use.</p>
<h2>Official QuantFunc links</h2>
<ul>
<li><a href="https://huggingface.co/QuantFunc" rel="ugc">Hugging Face — QuantFunc</a></li>
<li><a href="https://www.modelscope.cn/profile/QuantFunc" rel="ugc">ModelScope — QuantFunc</a></li>
<li><a href="https://discord.gg/jCp9TpFWcn" rel="ugc">Official Discord</a></li>
<li><a href="https://www.quantfunc.com/" rel="ugc">QuantFunc website</a></li>
<li><a href="https://github.com/QuantFunc/ComfyUI-QuantFunc" rel="ugc">ComfyUI-QuantFunc plugin and workflows</a></li>
</ul>
<h2>License terms</h2>
<p><strong>Built with Qwen.</strong> These derived weights retain the Qwen Research License: the community grant is for research or evaluation purposes only. Commercial use requires a separate license from the original licensor. Redistribution must include the original agreement, notices of modification, and the required Notice file. See the <a href="https://huggingface.co/Qwen/Qwen-Image-2.1/blob/main/LICENSE" rel="ugc">full license</a>.</p>
<h2>Source and weight integrity</h2>
<p>Original QuantFunc release: <a href="https://huggingface.co/QuantFunc/Qwen-Image-2.1-4bit" rel="ugc">QuantFunc/Qwen-Image-2.1-4bit</a>.</p>
<ul>
<li><code>qwen_image_21_base_quantfunc_int4_r128.safetensors</code> — SHA256: <code>920a3be0b5b559ea67e97586f68fae72aa0fd60ab97537ebdcfbf1a598634776</code></li>
<li><code>qwen_image_21_base_quantfunc_int4_r32.safetensors</code> — SHA256: <code>177a9b59a7d65ad41925cb3fc29d9a8a7b1deb5732d7aea1b36405f35b612653</code></li>
</ul>
<p>Showcase media is reproduced from the original QuantFunc model card. Exact seeds, prompts and rank variants are not supplied for every example. Performance figures are QuantFunc's reported measurements under the stated conditions; they are not guarantees for other workflows.</p>

QuantFunc A4W4 INT4 — Qwen-Image-2.1

<p><strong>True 4-bit inference — A4W4 (4-bit activations × 4-bit weights).</strong></p>
<p>QuantFunc's core quantized matrix multiplications use <strong>INT4 activations and INT4 weights</strong> on its INT4 inference backend. A4W4 describes the compute precision used during inference, alongside the reduced weight size and memory bandwidth demand.</p>
<p>QuantFunc INT4 r32 quantization of Qwen-Image-2.1. Smaller rank variant. Requires ComfyUI-QuantFunc or the QuantFunc inference engine.</p><p>File: <code>qwen_image_21_base_quantfunc_int4_r32.safetensors</code><br />Size: 4.33 GB (4.03 GiB)<br />SHA256: <code>177a9b59a7d65ad41925cb3fc29d9a8a7b1deb5732d7aea1b36405f35b612653</code></p><p>Research/evaluation use only under the Qwen Research License; commercial use requires a separate license. Built with Qwen.</p>

qwen_image_21_base_quantfunc_int4_r32.safetensors

quantfuncA4W4INT4Qwen_a4w4INT4R32.zip

A4W4 INT4 r32

A4W4 INT4 r128

<p><strong>True 4-bit inference — A4W4 (4-bit activations × 4-bit weights).</strong></p>
<p>QuantFunc's core quantized matrix multiplications use <strong>INT4 activations and INT4 weights</strong> on its INT4 inference backend. A4W4 describes the compute precision used during inference, alongside the reduced weight size and memory bandwidth demand.</p>
<p>QuantFunc INT4 r128 quantization of Qwen-Image-2.1. Quality-first variant. Requires ComfyUI-QuantFunc or the QuantFunc inference engine.</p><p>File: <code>qwen_image_21_base_quantfunc_int4_r128.safetensors</code><br />Size: 4.66 GB (4.34 GiB)<br />SHA256: <code>920a3be0b5b559ea67e97586f68fae72aa0fd60ab97537ebdcfbf1a598634776</code></p><p>Research/evaluation use only under the Qwen Research License; commercial use requires a separate license. Built with Qwen.</p>

Files with SHA256: 78fd9fecb374a571fbbfd2cee68ca3071d73cb63d0433ffcac99ec4db5bba917