Qwen3-VL-32B-Instruct-ultra-uncensored-heretic
32BQwenheretic
from Qwen/Qwen3-VL-32B-Instruct
↓ 56.2K♥ 307 repos
Qwen3-VL-32B-Instruct-abliterated
Qwen3-VL-32B-Instruct-abliterated is a refusal-ablated variant of Qwen/Qwen3-VL-32B-Instruct, published on Hugging Face by huihui-ai. It is 32B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 7 repositories in Transformers, GGUF, GGUF (imatrix) and FP8 formats, totalling 8.4K downloads.
Specification
- Model name
- Qwen3-VL-32B-Instruct-abliterated
- Source repository
- huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated
- Base model
- Qwen/Qwen3-VL-32B-Instruct
- Publisher
- huihui-ai
- Parameters
- 32B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, GGUF (imatrix), FP8, MLX
- First indexed
- 22 Oct 2025
- Last updated
- 22 Aug 2026
Ablation technique
How Qwen3-VL-32B-Instruct-abliterated was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“This is an uncensored version of Qwen/Qwen3-VL-32B-Instruct created with abliteration (see remove-refusals-with-transformers to know more about it).”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Ollama (publisher-documented)
ollama run huihui_ai/qwen3-vl-abliterated:32b-instructllama.cpp / GGUF
hf download mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-GGUF --local-dir ./qwen3-vl-32b-instruct-abliterated
llama-cli -m ./qwen3-vl-32b-instruct-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated --trust-remote-codeDownloads and variants (6)
Every published repository of Qwen3-VL-32B-Instruct-abliterated, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-GGUF | GGUF | 3.2K |
| huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliteratedsource | Transformers | 2.8K |
| mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-i1-GGUF | GGUF (imatrix) | 1.4K |
| Heouzen/Huihui-Qwen3-VL-32B-Instruct-FP8-abliterated | FP8 | 595 |
| divinetribe/Huihui-Qwen3-VL-32B-Instruct-abliterated-4bit-mlx | MLX | 413 |
| gsting/Qwen3-VL-32B-Instruct-abliterated | Transformers | 7 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3-VL-32B-Instruct-abliterated will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- Capability is inherited from Qwen/Qwen3-VL-32B-Instruct, not added by ablation — at 32B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is apache-2.0, but the terms that bind you are Qwen/Qwen3-VL-32B-Instruct’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for Qwen3-VL-32B-Instruct-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with huihui-ai, and does not independently verify publisher claims. See responsible use.