Instructions to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="terasut/typhoon-si-med-thinking-4b-research-preview-GGUF",
	filename="typhoon-si-med-thinking-4b-research-preview-BF16-GGUF.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use terasut/typhoon-si-med-thinking-4b-research-preview-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 terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

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 terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
# Run inference directly in the terminal:
./llama-cli -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

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 terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Use Docker

docker model run hf.co/terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

LM Studio
Jan

vLLM

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "terasut/typhoon-si-med-thinking-4b-research-preview-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": "terasut/typhoon-si-med-thinking-4b-research-preview-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Ollama
How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with Ollama:
```
ollama run hf.co/terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
```

Unsloth Studio

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for terasut/typhoon-si-med-thinking-4b-research-preview-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for terasut/typhoon-si-med-thinking-4b-research-preview-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for terasut/typhoon-si-med-thinking-4b-research-preview-GGUF to start chatting

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

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 "terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16" \
  --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"

Docker Model Runner
How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with Docker Model Runner:
```
docker model run hf.co/terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16
```

Lemonade

How to use terasut/typhoon-si-med-thinking-4b-research-preview-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Run and chat with the model

lemonade run user.typhoon-si-med-thinking-4b-research-preview-GGUF-BF16

List all available models

lemonade list

Hermes Agent

How to use terasut/typhoon-si-med-thinking-4b-research-preview-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 terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

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 terasut/typhoon-si-med-thinking-4b-research-preview-GGUF:BF16

Run Hermes

hermes

Atomic Chat

Typhoon-Si-Med-Thinking-4B (GGUF)

This repository provides GGUF conversions of Typhoon-Si-Med-Thinking-4B (Research Preview) for efficient local inference with llama.cpp and compatible runtimes.

Original model (HF): typhoon-ai/typhoon-si-med-thinking-4b-research-preview
Conversion source used: https://huggingface.co/typhoon-ai/typhoon-si-med-thinking-4b-research-preview
Converted with: llama.cpp/convert_hf_to_gguf.py (llama.cpp)

For full model details (training, benchmarks, prompt guidance, and intended use), please refer to the original model card: https://huggingface.co/typhoon-ai/typhoon-si-med-thinking-4b-research-preview

About the model

Typhoon-Si-Med-Thinking-4B is a 4B-parameter medical reasoning model built on the Qwen3 architecture and tuned to produce reasoned answers and (optionally) ranked lists of plausible candidates—closer to differential diagnosis style reasoning than “single best answer” outputs.

Provided GGUF files

File	Quantization	Notes
`*-BF16-GGUF.gguf`	BF16	Highest fidelity; larger RAM/VRAM usage.
`*-Q8_0-GGUF.gguf`	Q8_0	Smaller/faster; slight quality tradeoff vs BF16.

Check the Files and versions tab for exact filenames and sizes.

Conversion details (reproducibility)

This GGUF set was produced by converting the HF model to GGUF and then quantizing:

# 1) Convert HF -> GGUF (BF16)
python llama.cpp/convert_hf_to_gguf.py \
  typhoon-ai/typhoon-si-med-thinking-4b-research-preview \
  --outtype bf16 \
  --outfile typhoon-si-med-thinking-4b-research-preview-BF16-GGUF.gguf

# 2) Quantize (Q8_0)
python llama.cpp/convert_hf_to_gguf.py \
  typhoon-ai/typhoon-si-med-thinking-4b-research-preview \
  --outtype q8_0 \
  --outfile typhoon-si-med-thinking-4b-research-preview-Q8_0-GGUF.gguf

Model tree for terasut/typhoon-si-med-thinking-4b-research-preview-GGUF

Base model

Qwen/Qwen3-4B-Instruct-2507

Finetuned

typhoon-ai/typhoon-si-med-thinking-4b-research-preview

Quantized

(3)

this model

terasut
/

typhoon-si-med-thinking-4b-research-preview-GGUF

Typhoon-Si-Med-Thinking-4B (GGUF)

About the model

Provided GGUF files

Conversion details (reproducibility)

Links

Model tree for terasut/typhoon-si-med-thinking-4b-research-preview-GGUF