Instructions to use migarcoes/Qwen3.5-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use migarcoes/Qwen3.5-4B with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="migarcoes/Qwen3.5-4B",
	filename="Qwen3.5-4B-MIO-Q4_K_M.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use migarcoes/Qwen3.5-4B with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf migarcoes/Qwen3.5-4B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf migarcoes/Qwen3.5-4B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf migarcoes/Qwen3.5-4B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf migarcoes/Qwen3.5-4B: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 migarcoes/Qwen3.5-4B:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf migarcoes/Qwen3.5-4B: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 migarcoes/Qwen3.5-4B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf migarcoes/Qwen3.5-4B:Q4_K_M

Use Docker

docker model run hf.co/migarcoes/Qwen3.5-4B:Q4_K_M

LM Studio
Jan

vLLM

How to use migarcoes/Qwen3.5-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "migarcoes/Qwen3.5-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "migarcoes/Qwen3.5-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/migarcoes/Qwen3.5-4B:Q4_K_M

Ollama
How to use migarcoes/Qwen3.5-4B with Ollama:
```
ollama run hf.co/migarcoes/Qwen3.5-4B:Q4_K_M
```

Unsloth Studio

How to use migarcoes/Qwen3.5-4B 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 migarcoes/Qwen3.5-4B 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 migarcoes/Qwen3.5-4B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for migarcoes/Qwen3.5-4B to start chatting

How to use migarcoes/Qwen3.5-4B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf migarcoes/Qwen3.5-4B:Q4_K_M

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": "migarcoes/Qwen3.5-4B:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use migarcoes/Qwen3.5-4B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf migarcoes/Qwen3.5-4B: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 migarcoes/Qwen3.5-4B:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use migarcoes/Qwen3.5-4B with Docker Model Runner:
```
docker model run hf.co/migarcoes/Qwen3.5-4B:Q4_K_M
```

Lemonade

How to use migarcoes/Qwen3.5-4B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull migarcoes/Qwen3.5-4B:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.5-4B-Q4_K_M

List all available models

lemonade list

Qwen3.5-4B GGUF (Metadata Corrected)

Esta es una versión del modelo Qwen3.5-4B convertida y cuantizada a formato GGUF utilizando llama.cpp.

¿Por qué esta versión?

Generada localmente para garantizar la precisión de la metadata de los 4.21 B de parámetros. Este modelo representa el "punto dulce" para hardware con 8GB de RAM, ofreciendo una capacidad de razonamiento superior a los modelos menores manteniendo una velocidad usable.

Archivos incluidos

Q4_K_M: El balance ideal. Permite una ejecución estable sin agotar la RAM del sistema (~2.51 GiB).

Rendimiento (Benchmark Local)

Resultados obtenidos en un Intel Core i5-4460 @ 3.20GHz utilizando 4 hilos en CPU:

Model	Size	Params	Backend	Threads	Test	t/s
qwen35 4B Q4_K_M	2.51 GiB	4.21 B	CPU	4	pp512	22.22 ± 0.02
qwen35 4B Q4_K_M	2.51 GiB	4.21 B	CPU	4	tg128	4.86 ± 0.01

Notas de Hardware

En el i5-4460, este modelo genera texto a una velocidad de lectura humana (~5 tokens/s), lo que lo hace ideal para asistentes locales.

Instrucciones de uso

./llama-cli -m Qwen3.5-4B-MIO-Q4_K_M.gguf -p "Escribe un correo formal pidiendo vacaciones" -n 256

Downloads last month: 24

GGUF

Model size

4B params

Architecture

qwen35

Hardware compatibility

4-bit

8-bit

16-bit

Model tree for migarcoes/Qwen3.5-4B

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Quantized

(213)

this model