Instructions to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
# Run inference directly in the terminal:
llama cli -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
# Run inference directly in the terminal:
llama cli -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
# Run inference directly in the terminal:
./llama-cli -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
# Run inference directly in the terminal:
./build/bin/llama-cli -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Use Docker

docker model run hf.co/crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

LM Studio
Jan
Ollama
How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with Ollama:
```
ollama run hf.co/crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
```
Unsloth Desktop

How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with Docker Model Runner:
```
docker model run hf.co/crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P
```

Lemonade

How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Run and chat with the model

lemonade run user.Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP-Q2_K_P

List all available models

lemonade list

Hermes Agent

How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

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 crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P

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 "crotron/Qwen3.6-27B-Uncensored-HauhauCS-Balanced-MTP:Q2_K_P" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

Q2_K_P... yes please

by sa-ai - opened Jun 14

Discussion

sa-ai

Jun 14

Your Q4_K_P model is in my top 5 tested recently (past 2 months <=40B), thank you for your efforts.

The original balanced HauhauCS Q2_K_P was showing strong results too, so it carries weight as well... especially since users can run it on a single 16 gig card at 64k+ ctx. MTP would be the cherry on top. This should allow for >30t/s on a 5060 ti.

Thanks again

julitroalves

Jun 24

Me either, for me is the best option for coding, it's very smart.

sa-ai

Jun 25

•

edited Jun 25

for anyone else interested... I was able to pull this off (mtp for 27B, any Q from 3.5 or 3.6 Qwen) using 27B_MTP.gguf and convert.py from https://huggingface.co/havenoammo/Qwen3.6-27B-MTP-UD-GGUF
Went from 33 to 56t/s (llama.cpp: --spec-type draft-mtp --spec-draft-n-max 2) on dual 3090s w/nvlink. getting about 50t/s on dual 5060ti's

He has one also for the 35B

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment