Instructions to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF 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 heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "heack/Math-Physics-Qwen3.6-10.5B-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "heack/Math-Physics-Qwen3.6-10.5B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Ollama
How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with Ollama:
```
ollama run hf.co/heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

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": "heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with Docker Model Runner:
```
docker model run hf.co/heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M
```

Lemonade

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Math-Physics-Qwen3.6-10.5B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

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 heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use heack/Math-Physics-Qwen3.6-10.5B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M

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 "heack/Math-Physics-Qwen3.6-10.5B-GGUF:Q4_K_M" \
  --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"

Math & Physics 10.5B

A text model for mathematics and physics, pruned and distilled from Qwen3.6-35B-A3B. It has 10.5 billion parameters and fits the tested RTX 5060 Ti 16 GB.

Download	Size	Choose this for
Q4_K_M	6.48 GB	Smaller download and faster generation
Q8_0	11.17 GB	The version used for the accuracy results below

This release accepts text. It does not accept pictures.

Run it on Windows

Requirements: Windows x64, Python 3.10+, an NVIDIA GPU and a current NVIDIA driver. Tested on an RTX 5060 Ti 16 GB.

1. Download the model and examples in PowerShell:

py -m venv .venv
.\.venv\Scripts\python.exe -m pip install --upgrade huggingface_hub
.\.venv\Scripts\hf.exe download heack/Math-Physics-Qwen3.6-10.5B-GGUF g64full_kd-Q4_K_M.gguf examples/chat.py setup-runtime.ps1 --local-dir math-physics
cd math-physics

2. Download the tested llama.cpp runtime:

Run this in the same PowerShell window. It downloads and checks the exact Windows CUDA runtime used by these instructions:

powershell -NoProfile -ExecutionPolicy Bypass -File .\setup-runtime.ps1

3. Start the model:

.\runtime\llama-server.exe -m .\g64full_kd-Q4_K_M.gguf -ngl 99 --host 127.0.0.1 --port 18085 -c 8192 -np 1 --jinja --reasoning-format deepseek -fa on -ctk q8_0 -ctv q8_0 -a math-physics-10.5b

Wait until the server is ready, then open http://127.0.0.1:18085 in your browser and ask a question.

4. Check the API: open another PowerShell window in math-physics and run:

py examples/chat.py --url http://127.0.0.1:18085

The example checks 3x + 7 = 22 and Newton's law for a 2 kg object accelerating at 3 m/s². Expected answers: x = 5 and 6 N. It saves verification.json. Both Q4 and Q8 passed this check on the tested setup.

For Q8, download g64full_kd-Q8_0.gguf instead and change the filename in step 3. No separate base model is needed.

Results

Q8 results on small fixed subsets:

Test	Correct
MATH-500, levels 4–5	32/40 (80%)
MMLU-Pro physics	19/30 (63.3%)
UGPhysics	11/30 (36.7%)
GPQA-Diamond physics	24/30 (80%)

These are screening results, not full-dataset scores or proof of matching the original Qwen model. Q4 accuracy has not been measured in this suite. Answers can be incorrect; check important calculations.

Evaluation · Q4 usage verification · Q8 usage verification · Technical details

Online API

No hosted endpoint is available yet. Hugging Face rejected this account's ZeroGPU deployment request with a PRO subscription requirement. The commands above run the model locally.

License

Noncommercial use: CC BY-NC-SA 4.0. Commercial use: US$1,000 once for a perpetual license to this model's purchased version. Each model is licensed separately. Hosting and API usage are not included. Upstream licenses and prior valid rights remain in effect.

Select USD and include heack/Math-Physics-Qwen3.6-10.5B-GGUF, your Hugging Face username, and the licensee's name with your payment. Read the commercial terms.

Downloads last month: 390

GGUF

Model size

10B params

Architecture

qwen35moe

Hardware compatibility

4-bit

8-bit

Model tree for heack/Math-Physics-Qwen3.6-10.5B-GGUF

Base model

Qwen/Qwen3.6-35B-A3B

Quantized

(965)

this model