NeuralNet-Hub

Qwen3.5-9B-Uncensored-W4A16

README

License: apache-2.0

🌟 Qwen3.5-9B Uncensored NVFP4 Quantization by NeuralNet 🧠🤖

This is an W4A16-quantized version of Qwen/Qwen3.5-9B, produced through a combination of fine-tuning and abliteration. This model was obtained by quantizing the weights of Qwen/Qwen3.5-9B to INT4 data type while keeping activations in original precision, ready for inference with vLLM The reduction is less than the theoretical 75% because the vision encoder, token embeddings, and linear attention layers remain in BF16.

Only the weights of the linear operators within transformer blocks are quantized using LLM Compressor. The vision encoder, token embeddings, and linear attention layers are not quantized.

[!IMPORTANT] This format is not compatible with Ampere, Ada Lovelace, Blackwell or Hopper GPUs.

🔓 No Filters. No Limits. Just Answers.

This model powers UncensoredGPT

Ask anything. Get real answers. No restrictions.

Most AI models are trained to refuse. They hedge, they deflect, they lecture. UncensoredGPT is built on the opposite philosophy: that access to information should be unrestricted, and that adults are capable of deciding what they need to know.

This model is the engine behind UncensoredGPT, a platform providing unfiltered, honest responses for cybersecurity, education, content creation, research, or straightforward conversation. The refusals and content restrictions present in the original Qwen3.5-9B have been removed through a combination of supervised fine-tuning and abliteration, resulting in a model that responds directly across topics that standard models typically refuse.

Why stay in the system when you can have unrestricted answers, privacy by default, and complete freedom of information?

Ready to experience the freedom of unrestricted AI? Join the waitlist at uncensoredgpt.ai — limited spots available.

⚡ Deployment with vLLM

This quantized model is intended to be served using vLLM (vllm>=0.21.0 recommended).

Quick Start

bash
vllm serve NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16 \
  --dtype bfloat16 \
  --kv-cache-dtype fp8 \
  --max-model-len 262144 \
  --reasoning-parser qwen3 \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_coder

Using a Config File

yaml
# Deploy with: vllm serve --config config.yaml
# Optimized for NVIDIA RTX 6000 PRO (Blackwell)
# Benchmarked: ~85-90 parallel requests, up to 1000 tok/sec at higher context lengths

model: NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16
dtype: bfloat16
kv-cache-dtype: fp8
gpu-memory-utilization: 0.95
max-model-len: 262144
max-num-batched-tokens: 4096
max-num-seqs: 200
max-cudagraph-capture-size: 209
enable-prefix-caching: true
trust-remote-code: true

reasoning-parser: qwen3
enable-auto-tool-choice: true
tool-call-parser: qwen3_coder

default-chat-template-kwargs: '{"enable_thinking": false}'

download-dir: /workspace/models
host: 0.0.0.0
port: 18000

bash
vllm serve --config config.yaml

💬 Chat API Usage

Qwen3.6 uses a standard chat template compatible with OpenAI-format APIs. Thinking mode is enabled by default.

Thinking Mode (Default)

python
from openai import OpenAI

client = OpenAI(base_url="http://localhost:18000/v1", api_key="EMPTY")

messages = [{"role": "user", "content": "Your message here"}]

response = client.chat.completions.create(
    model="NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16",
    messages=messages,
    max_tokens=32768,
    temperature=1.0,
    top_p=0.95,
    extra_body={"top_k": 20},
)
print(response.choices[0].message.content)

Non-Thinking (Instruct) Mode

python
response = client.chat.completions.create(
    model="NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16",
    messages=messages,
    max_tokens=8192,
    temperature=0.7,
    top_p=0.8,
    presence_penalty=1.5,
    extra_body={
        "top_k": 20,
        "chat_template_kwargs": {"enable_thinking": False},
    },
)

Image Input

python
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}},
            {"type": "text", "text": "Describe this image in detail."}
        ]
    }
]

response = client.chat.completions.create(
    model="NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16",
    messages=messages,
    max_tokens=32768,
    temperature=1.0,
    top_p=0.95,
    extra_body={"top_k": 20},
)

⚙️ Recommended Sampling Parameters

Table with columns: Mode, temperature, top_p, top_k, presence_penalty
Mode	temperature	top_p	top_k	presence_penalty
Thinking — general tasks	1.0	0.95	20	0.0
Thinking — precise coding	0.6	0.95	20	0.0
Instruct (non-thinking)	0.7	0.80	20	1.5

📥 Download with huggingface-cli

Install the CLI

bash
pip install -U "huggingface_hub[cli]"

Download the Full Repository

bash
huggingface-cli download NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16 --local-dir ./Qwen3.5-9B-Uncensored-W4A16

Download Specific Files

bash
huggingface-cli download NeuralNet-Hub/Qwen3.5-9B-Uncensored-W4A16 \
  --include "*.safetensors" \
  --local-dir ./Qwen3.5-9B-Uncensored-W4A16

🌐 Contact Us

NeuralNet is a pioneering AI solutions provider that empowers businesses to harness the power of artificial intelligence.

Website: https://neuralnet.solutions Email: info[at]neuralnet.solutions

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider