Instructions to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

Use Docker

docker model run hf.co/SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SandLogicTechnologies/Qwen3-4B-Thinking-2507-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": "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

SGLang

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with Ollama:
```
ollama run hf.co/SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
```
Unsloth Desktop

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-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": "SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with Docker Model Runner:
```
docker model run hf.co/SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M
```

Lemonade

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3-4B-Thinking-2507-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-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 SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507-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 "SandLogicTechnologies/Qwen3-4B-Thinking-2507-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"

Quantized Qwen3-4B-Thinking-2507

This repository provides the Q4 and Q5 quantized version of the Qwen3-4B-Thinking-2507 model. This model emphasizes depth of reasoning (“thinking mode”). It outputs an internal “thought” component (within / boundaries) followed by the final generated content. This model Significantly improved reasoning and general capabilities—including logic, math, science, coding, instruction following, tool usage, text generation, alignment with human preferences, and 256K long-context understanding. These quantized models will able to run on the CPU's and edge devices and can utilize the model capabality without limitation of the hardware.

Model Overview

Original Model: Qwen3-4B-Thinking-2507
Thinking Mode: Default enabled; tag not required
Architecture: decoder-only model
Base Model: Qwen3-4B-Thinking-2507
Quantized Version:
- Q4_K_M
- Q5_K_M
Modalities: Text
Developer: Qwen
License: Apache 2.0 License
Languages: English

Quantization Details

Q4_K_M Version

Approx. ~71% size reduction
Lower memory footprint (~2.33 GB)
Best suited for deployment on edge devices or low-resource GPUs
Slight performance degradation in complex reasoning scenarios

Q5_K_M Version

Approx. ~66% size reduction
Lower memory footprint (~2.69 GB)
Better performance retention, recommended when quality is a priority

Key Features

Significantly improved reasoning: logical reasoning, mathematics, science, coding, and academic benchmarks.
Better general capabilities: instruction following, tool usage, text generation, and alignment with human preferences.
Long-Context Understanding: Can handle up to 256K tokens, enabling analysis of very large documents.
Large-Scale Transformer: 36-layer decoder-only Transformer with Grouped Query Attention for efficient computation..
Deployment Ready: Compatible with CPUs, lightweight GPUs, and frameworks like vllm, llama.cpp.

Usage Example

Using llama.cpp for inference:

./llama-cli -hf SandLogicTechnologies/Qwen3-4B-Thinking-2507 -p "Give me a short introduction to large language model."

Recommended Use Cases

Advanced Reasoning Tasks: Logical reasoning, mathematics, science, and coding problems.
Academic Assistance: Solving benchmark questions, research summarization, and educational content generation.
Instruction Following: Chatbots and virtual assistants that respond accurately to user instruction.
Long-Context Applications: Analyzing and generating content from very large documents (up to 256K tokens).
Deployment in Low-Resource Environments: Running efficiently on CPUs, edge devices, or lightweight GPUs.

Acknowledgments

These quantized models are based on the original work by the Qwen development team.

Special thanks to:

The Qwen team for developing and releasing the Qwen3-4B-Thinking-2507 model.
Georgi Gerganov and the entire llama.cpp open-source community for enabling efficient model quantization and inference via the GGUF format.

Contact

For any inquiries or support, please contact us at [email protected] or visit our Website.

Downloads last month: 53

GGUF

Model size

4B params

Architecture

qwen3

Hardware compatibility

4-bit

5-bit

Model tree for SandLogicTechnologies/Qwen3-4B-Thinking-2507-GGUF

Base model

Qwen/Qwen3-4B-Thinking-2507

Quantized

(145)

this model