Huihui Qwen3 VL 32B Instruct Abliterated needs about 22.4 GB of memory at Q4_K_M (a 19.8 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 11 quants, from 12.3 GB (Q2_K) to 34.8 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
Huihui Qwen3 VL 32B Instruct Abliterated – a 32.8B vision‑language model
This model is a 32.8B parameter Qwen3‑VL variant that has been abliterated to remove refusal behaviour. It supports vision tasks and is available in multiple GGUF quantisations. The model is licensed under apache‑2.0.
Good at
- Vision‑language inference
- Large context handling
- Uncensored text generation
Limits
- No safety filtering – use with caution
- Requires large VRAM for higher quants
- No default vision safety checks
Best for
- Research on vision‑language models
- Controlled content generation
- Image‑prompted dialogue
Who it’s for. Researchers and developers needing a high‑capacity vision‑language model for experimentation. Use in environments where output review is possible and safety is not a primary concern.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 12.3 GB | 15.0 GB | Too big | Get file → | |
| Q3_K_S | 14.4 GB | 17.0 GB | Too big | Get file → | |
| Q3_K_M | 16.0 GB | 18.6 GB | Too big | Get file → | |
| Q3_K_L | 17.3 GB | 20.0 GB | Too big | Get file → | |
| IQ4_XS | 17.9 GB | 20.5 GB | Too big | Get file → | |
| Q4_K_S | 18.8 GB | 21.4 GB | Too big | Get file → | |
| Q4_K_M | 19.8 GB | 22.4 GB | Too big | Get file → | |
| Q5_K_S | 22.6 GB | 25.3 GB | Too big | Get file → | |
| Q5_K_M | 23.2 GB | 25.9 GB | Too big | Get file → | |
| Q6_K | 26.9 GB | 29.5 GB | Too big | Get file → | |
| Q8_0 | 34.8 GB | 37.5 GB | Too big | Get file → |
For 8 GB VRAM, Q2_K (12.34 GB) or Q3_K_S (14.39 GB) are suitable. For 16 GB VRAM, Q4_K_M (19.76 GB) or Q5_K_S (22.64 GB) provide a balance of size and quality. Larger VRAM allows Q6_K or Q8_0 for best performance.
What GPU do you need for Huihui Qwen3 VL 32B Instruct Abliterated?
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: the best quant that fits is Q2_K (12.3 GB file, about 15.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q4_K_M (19.8 GB file, about 22.4 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (26.9 GB file, about 29.5 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (34.8 GB file, about 37.5 GB in use).
Which Huihui Qwen3 VL 32B Instruct Abliterated quant should you choose?
Q4_K_M is the usual starting point for Huihui Qwen3 VL 32B Instruct Abliterated: it balances file size and quality well. Here it is a 19.8 GB file.
The Q8_0 file is 34.8 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 Q2_K at 12.3 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about Huihui Qwen3 VL 32B Instruct Abliterated
How much VRAM does Huihui Qwen3 VL 32B Instruct Abliterated need?
About 22.4 GB at Q4_K_M with an 8K context window. That is the 19.8 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 Huihui Qwen3 VL 32B Instruct Abliterated on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 15.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Huihui Qwen3 VL 32B Instruct Abliterated GGUF?
It depends on the quant: Q2_K is 12.3 GB, Q3_K_S is 14.4 GB, Q3_K_M is 16.0 GB, Q3_K_L is 17.3 GB, IQ4_XS is 17.9 GB, Q4_K_S is 18.8 GB and 5 more.
What context length does Huihui Qwen3 VL 32B Instruct Abliterated support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.
Where can I download Huihui Qwen3 VL 32B Instruct Abliterated GGUF?
The files are on Hugging Face in the mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-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 Huihui Qwen3 VL 32B Instruct Abliterated 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 this model on 8 GB VRAM?
Yes, using Q2_K or Q3_K_S quantisations will fit within 8 GB VRAM with approximate memory usage of 15–17 GB at 8k context.
Does this model support image input?
Yes, it detects vision abilities and can process images in prompts.
Is there any safety filtering?
The model has reduced safety filtering; it may produce sensitive or inappropriate content and should be monitored.