Instructions to use dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK with MLX:

# Make sure mlx-vlm is installed
# pip install --upgrade mlx-vlm

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK")
config = load_config("dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK

Run Hermes

hermes

OpenClaw new

How to use dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Important: This model uses the JANG quantization format — the GGUF equivalent for MLX on Apple Silicon. Currently only supported by MLX Studio and the jang-tools Python package.

MLX Studio — the only app that natively supports JANG models

Qwen 3.5 VL 397B — JANG_2L + CRACK

JANG mixed-precision · CRACK abliterated · Vision-Language · No guardrails · 187 GB

What Is This?

This is Qwen 3.5 VL 397B — a 397B parameter hybrid SSM/Attention Mixture-of-Experts model with 512 experts (10 active per token), and built-in vision.

It has been:

JANG quantized — JANG_2L profile (8-bit attention, 6-bit important, 2-3-bit experts) — 187 GB
CRACK abliterated — permanent weight-level removal of safety refusal


Architecture	Qwen 3.5 VL MoE — 397B total, ~17B active, 512 experts, hybrid SSM/FA
Quantization	JANG_2L (8/6/3/2-bit mixed, 3.72 avg) — 187 GB
HarmBench	98.4% (315/320)
Compliance	8/8
Vision	Yes — via MLX Studio / vMLX
Thinking	ON/OFF supported
MMLU	86.5% (180/208)
Speed	~33 tok/s (M4 Ultra 256GB)
Fits on	256 GB Macs

Also see: JANG_1L version — 112 GB, 96.2% HarmBench (fits on 128 GB Macs)

HarmBench Results

315/320 (98.4%)

Category	Score
Chemical / Biological	42/42	100%
Copyright	80/80	100%
Cybercrime / Intrusion	52/52	100%
Harmful	18/18	100%
Misinformation / Disinfo	53/54	98%
Illegal	51/53	96%
Harassment / Bullying	19/21	90%

Install & Usage

pip install "jang[mlx]"

from jang_tools.loader import load_jang_model
from mlx_lm import generate

model, tokenizer = load_jang_model("dealignai/Qwen3.5-VL-397B-A17B-JANG_2L-CRACK")

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

response = generate(model, tokenizer, prompt=prompt, max_tokens=2000)
print(response)

Thinking Mode

Thinking is ON by default. To disable:

prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True,
    enable_thinking=False, tokenize=False)

About JANG

JANG (Jang Adaptive N-bit Grading) is a mixed-precision quantization format for Apple Silicon — the GGUF equivalent for MLX.

About CRACK

CRACK (Controlled Refusal Ablation via Calibrated Knockouts) removes safety alignment from LLMs at the weight level using per-layer projected vectors from structurally-mirrored prompt pairs.

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