Qwen3-Vl-235B-A22B-Instruct needs about 144.2 GB of memory at Q4_K_M (a 142.2 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 25 quants, from 54.4 GB (TQ1_0) to 470.3 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 6.3 GB.
Qwen3-VL‑235B‑A22B‑Instruct: 235 B vision‑language model with 22 B active MoE parameters
Qwen3‑VL‑235B‑A22B‑Instruct is a 235 B parameter model with a 22 B active mixture of experts. It supports vision tasks and has a native context length of 262 144 tokens. The model is released under the Apache‑2.0 license.
Good at
- Vision‑language reasoning
- Long‑context text and video understanding
- Multimodal agent interaction
Limits
- Requires large VRAM for full model
- No native vision support in older GGUFs
- License requires attribution
Best for
- Developing visual agents for GUIs
- Generating code from images
- Processing long documents with images
Who it’s for. Developers building multimodal applications that need high‑accuracy vision and text reasoning. Researchers requiring a large MoE model for experimentation.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| TQ1_0 | 54.4 GB | 56.5 GB | Too big | Get file → | |
| IQ1_S | 62.8 GB (2 parts) | 64.9 GB | Too big | Get file → | |
| IQ1_M | 69.6 GB (2 parts) | 71.7 GB | Too big | Get file → | |
| Q2_K | 85.7 GB (2 parts) | 87.8 GB | Too big | Get file → | |
| Q2_K_L | 85.8 GB (2 parts) | 87.9 GB | Too big | Get file → | |
| Q2_K_XL | 88.0 GB (2 parts) | 90.1 GB | Too big | Get file → | |
| IQ3_XXS | 94.6 GB (2 parts) | 96.7 GB | Too big | Get file → | |
| Q3_K_S | 101.4 GB (3 parts) | 103.5 GB | Too big | Get file → | |
| Q3_K_XL | 103.7 GB (3 parts) | 105.8 GB | Too big | Get file → | |
| Q3_K_M | 112.4 GB (3 parts) | 114.5 GB | Too big | Get file → | |
| IQ4_XS | 125.3 GB (3 parts) | 127.4 GB | Too big | Get file → | |
| IQ4_NL | 132.6 GB (3 parts) | 134.7 GB | Too big | Get file → | |
| Q4_0 | 133.1 GB (3 parts) | 135.2 GB | Too big | Get file → | |
| Q4_K_S | 133.7 GB (3 parts) | 135.8 GB | Too big | Get file → | |
| Q4_K_XL | 134.1 GB (3 parts) | 136.2 GB | Too big | Get file → | |
| Q4_K_M | 142.2 GB (3 parts) | 144.2 GB | Too big | Get file → | |
| Q4_1 | 147.2 GB (3 parts) | 149.3 GB | Too big | Get file → | |
| Q5_K_S | 161.9 GB (4 parts) | 164.0 GB | Too big | Get file → | |
| Q5_K_M | 166.8 GB (4 parts) | 168.9 GB | Too big | Get file → | |
| Q5_K_XL | 166.9 GB (4 parts) | 169.0 GB | Too big | Get file → | |
| Q6_K | 193.0 GB (4 parts) | 195.1 GB | Too big | Get file → | |
| Q6_K_XL | 200.4 GB (5 parts) | 202.5 GB | Too big | Get file → | |
| Q8_0 | 249.9 GB (6 parts) | 252.0 GB | Too big | Get file → | |
| Q8_K_XL | 269.7 GB (6 parts) | 271.8 GB | Too big | Get file → | |
| BF16 | 470.3 GB (10 parts) | 472.4 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_0 (133 GB file, ~135 GB memory). For 12 GB VRAM, Q4_K_S (133 GB file, ~136 GB memory). For 16 GB VRAM, Q4_K_M (142 GB file, ~144 GB memory).
What GPU do you need for Qwen3-Vl-235B-A22B-Instruct?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which Qwen3-Vl-235B-A22B-Instruct quant should you choose?
Q4_K_M is the usual starting point for Qwen3-Vl-235B-A22B-Instruct: it balances file size and quality well. Here it is a 142.2 GB file.
The Q8_0 file is 249.9 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is TQ1_0 at 54.4 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3-Vl-235B-A22B-Instruct
How much VRAM does Qwen3-Vl-235B-A22B-Instruct need?
About 144.2 GB at Q4_K_M with an 8K context window. That is the 142.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Qwen3-Vl-235B-A22B-Instruct on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 56.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3-Vl-235B-A22B-Instruct GGUF?
It depends on the quant: TQ1_0 is 54.4 GB, IQ1_S is 62.8 GB, IQ1_M is 69.6 GB, Q2_K is 85.7 GB, Q2_K_L is 85.8 GB, Q2_K_XL is 88.0 GB and 19 more.
What context length does Qwen3-Vl-235B-A22B-Instruct support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.6 GB at 8K and 6.3 GB at 32K.
Where can I download Qwen3-Vl-235B-A22B-Instruct GGUF?
The files are on Hugging Face in the unsloth/Qwen3-VL-235B-A22B-Instruct-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Qwen3-Vl-235B-A22B-Instruct use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Qwen3‑VL‑235B‑A22B‑Instruct on 8 GB VRAM?
Use the Q4_0 quantized file; it requires about 135 GB of VRAM at 8 k context, which is beyond 8 GB. A smaller quant or a lower‑resolution model is needed.
Does this model support vision tasks?
Yes, the model has detected vision abilities and can process images and video.
What license does the model use?
The model is released under the Apache‑2.0 license.