CoolFace
Modelpublic

dealignai/Qwen3.5-VL-122B-A10B-UNCENSORED-JANG_2S

sourceHugging Faceapache-2.0updated 20d agoView on Hugging Face
4likes883downloads
Model Card
CRITICAL FIX (2026-03-21): Fixed chat_template.jinja — previous versions may have had thinking loop issues. Re-download if you downloaded before today.
Important: This model uses the JANG quantization format — the GGUF equivalent for MLX on Apple Silicon. Currently only supported by [MLX Studio](https://mlx.studio) and the jang-tools Python package. LM Studio, Ollama, and other apps do not support JANG yet.

<p align="center"> <a href="https://mlx.studio"><img src="https://raw.githubusercontent.com/jjang-ai/jangq/main/assets/mlx-studio-light.png" alt="MLX Studio" width="500"></a> </p>

<p align="center"> <a href="https://mlx.studio"><img src="https://mlx.studio/assets/screenshots/mlx-studio-featured.png?v=1" alt="MLX Studio App" width="600"></a> </p>

<h4 align="center"><a href="https://mlx.studio">MLX Studio</a> — the only app that natively supports JANG models</h4>

<div align="center">

<img src="dealign_mascot.png" width="128" />

<p align="center"> <a href="https://vmlx.net"><img src="vmlx-app.png" alt="vMLX — run JANG models on Apple Silicon" width="820"></a> </p>

<h3 align="center">⚡ All JANG models are meant to be run in <a href="https://vmlx.net">vMLX</a></h3>

Qwen 3.5 VL 122B-A10B — JANG_2S + CRACK

JANG mixed-precision · CRACK abliterated · No guardrails · VLM · 35 GB

<a href="https://ko-fi.com/jangq"><img src="https://img.shields.io/badge/Ko--fi-Support_Development-FF5E5B?logo=ko-fi&logoColor=white&style=for-the-badge" alt="Ko-fi"></a>

</div>

What Is This?

This is Qwen 3.5 122B-A10B — a 122B parameter Mixture-of-Experts model with 256 experts (8 active per token), hybrid GatedDeltaNet SSM + full attention architecture, and built-in vision-language capabilities.

It has been:

  1. 1.JANG quantized — JANG_2S profile (6-bit attention, 4-bit embeddings, 2-bit experts) — 35 GB, fits on 48 GB Macs
  2. 2.CRACK abliterated — permanent weight-level removal of safety refusal behavior

JANG's mixed-precision approach keeps attention weights at 6-bit (CRITICAL tier) while compressing MoE expert weights to 2-bit. On MoE models, CRITICAL is <5% of parameters — the quality boost from 6-bit attention is nearly free.

ArchitectureQwen 3.5 MoE — 122B total, 10B active, 256 experts
QuantizationJANG_2S (6/4/2-bit mixed) — 35 GB
AbliterationCRACK — permanent weight modification
VisionBuilt-in VLM (333 vision encoder tensors)
ThinkingSupports enable_thinking ON/OFF
Speed~51 tok/s on MacBook Pro M4 Max 128 GB
Fits on48 GB+ Macs

HarmBench Results (320 prompts)

CategoryScoreRate
Harmful content18/18100%
Copyright79/8099%
Misinformation52/5496%
Cybercrime & intrusion49/5294%
Harassment & bullying19/2190%
Chemical & biological36/4286%
Illegal activities39/5374%
Overall292/32091.2%

MMLU-200 Results (Per Subject)

This Model (JANG_2S + CRACK) vs Base Models

SubjectJANG_2S CRACKJANG_2S BaseMLX 2-bitJANG_4K BaseMLX 4-bit
35 GB38 GB36 GB69 GB64 GB
Abstract Algebra12/209/209/2016/2015/20
Anatomy15/2018/2011/2019/2018/20
Astronomy20/2020/2016/2019/2019/20
College CS14/2014/208/2015/2015/20
College Physics12/2015/2010/2014/2014/20
HS Biology18/2019/2015/2019/2019/20
HS Chemistry17/2018/2013/2018/2018/20
HS Mathematics11/2011/204/2014/2014/20
Logical Fallacies17/2016/2013/2019/2019/20
World Religions19/2018/2014/2019/2019/20
Total155/200158/200113/200172/200170/200
Accuracy77.5%79%56.5%86%85%

Key takeaways:

  • —CRACK surgery costs only 1.5 MMLU points vs unmodified JANG_2S (77.5% vs 79%)
  • —JANG_2S is 22.5 points better than MLX uniform 2-bit (79% vs 56.5%)
  • —Even CRACK'd, this model beats MLX 2-bit by 21 points (77.5% vs 56.5%)

Install & Usage

bash
pip install "jang[mlx]"
python
from jang_tools.loader import load_jang_model
from mlx_lm import generate

model, tokenizer = load_jang_model("dealignai/Qwen3.5-VL-122B-A10B-JANG_2S-CRACK")

messages = [{"role": "user", "content": "Your prompt here"}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True,
    enable_thinking=False, tokenize=False)

response = generate(model, tokenizer, prompt=prompt, max_tokens=500)
print(response)

VLM Inference

bash
pip install "jang[vlm]"
python
from jang_tools.loader import load_jang_vlm_model
from mlx_vlm import generate

model, processor = load_jang_vlm_model("dealignai/Qwen3.5-VL-122B-A10B-JANG_2S-CRACK")
result = generate(model, processor, "Describe this image.", image=["photo.jpg"], max_tokens=200)
print(result.text)

About JANG

JANG (Jang Adaptive N-bit Grading) is a mixed-precision quantization format for Apple Silicon — the GGUF equivalent for MLX. Instead of quantizing all weights at the same bit width, JANG classifies tensors into sensitivity tiers:

  • —CRITICAL (attention, routers, output head): 6-8 bit
  • —IMPORTANT (embeddings, linear attention): 4-6 bit
  • —COMPRESS (MLP/FFN, MoE experts): 2-3 bit

On MoE models where CRITICAL is <5% of parameters, this gives dramatically better quality than uniform quantization at the same size.

About CRACK

CRACK (Controlled Refusal Ablation via Calibrated Knockouts) removes safety alignment from LLMs at the weight level. No custom model files, no runtime hooks — the modification is permanent and runs at full native speed.

Links

<p align="center">

<a href="https://ko-fi.com/jangq"><img src="https://img.shields.io/badge/Ko--fi-SupportDevelopment-FF5E5B?logo=ko-fi&logoColor=white&style=flat-square" alt="Ko-fi"></a> <a href="https://x.com/dealignai"><img src="https://img.shields.io/badge/X-@dealignai-000000?logo=x&logoColor=white&style=flat-square" alt="X/Twitter"></a> <a href="https://github.com/jjang-ai/jangq"><img src="https://img.shields.io/badge/GitHub-jjang--ai/jangq-181717?logo=github&logoColor=white&style=flat-square" alt="GitHub"></a> <a href="https://mlx.studio"><img src="https://img.shields.io/badge/MLXStudio-App-blue?style=flat-square" alt="MLX Studio"></a> <a href="https://jangq.ai"><img src="https://img.shields.io/badge/Website-jangq.ai-green?style=flat-square" alt="Website"></a>

</p>

Disclaimer

This model is provided for research and educational purposes. The creators are not responsible for any misuse. By downloading this model, you agree to use it responsibly and in compliance with applicable laws.

한국어

Qwen 3.5 VL 122B — JANG_2S + CRACK

JANG 혼합정밀도 양자화 + CRACK 안전장치 제거 모델입니다.

항목내용
크기35 GB
MMLU77.5%
HarmBench91.2% 준수
최소 요구사양48 GB 메모리 Mac
bash
pip install "jang[mlx]"

GitHub · HuggingFace · MLX Studio · Ko-fi · X @dealignai

<p align="center">Created by <a href="https://jangq.ai">Jinho Jang</a> · 장진호 제작</p>