Instructions to use constructai/VibeThinker-3B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use constructai/VibeThinker-3B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="constructai/VibeThinker-3B-GGUF",
	filename="VibeThinker-3B-GGUF-F16.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 constructai/VibeThinker-3B-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 constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
llama cli -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
llama cli -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

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 constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
./llama-cli -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

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 constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
./build/bin/llama-cli -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

Use Docker

docker model run hf.co/constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

LM Studio
Jan

vLLM

How to use constructai/VibeThinker-3B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "constructai/VibeThinker-3B-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": "constructai/VibeThinker-3B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

Ollama
How to use constructai/VibeThinker-3B-GGUF with Ollama:
```
ollama run hf.co/constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
```

Unsloth Studio

How to use constructai/VibeThinker-3B-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 constructai/VibeThinker-3B-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 constructai/VibeThinker-3B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for constructai/VibeThinker-3B-GGUF to start chatting

How to use constructai/VibeThinker-3B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

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": "constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use constructai/VibeThinker-3B-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 constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

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 constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use constructai/VibeThinker-3B-GGUF with Docker Model Runner:
```
docker model run hf.co/constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL
```

Lemonade

How to use constructai/VibeThinker-3B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull constructai/VibeThinker-3B-GGUF:UD-Q4_K_XL

Run and chat with the model

lemonade run user.VibeThinker-3B-GGUF-UD-Q4_K_XL

List all available models

lemonade list

constructai/VibeThinker-3B-GGUF

This is a quantized version of the original VibeThinker-3B , converted to the GGUF format for efficient CPU/GPU inference with llama.cpp, Ollama, or any GGUF‑compatible runner.

Original Model

Author(s): WeiboAI
Source: VibeThinker-3B
Original License: MIT

Available Quantizations

Choose the quantization that fits your needs:

Quantization	File Size
`UD-IQ1_S`	791 MB
`UD-IQ1_M`	850 MB
`UD-IQ2_XXS`	948 MB
`Q2_K`	1.27 GB
`UD-IQ2_M`	1.14 GB
`UD-Q2_K_XL`	1.27 GB
`UD-IQ3_XXS`	1.28 GB
`Q3_K_S`	1.45 GB
`UD-IQ3_S`	1.46 GB
`Q3_K_M`	1.59 GB
`UD-Q3_K_M`	1.59 GB
`UD-Q3_K_XL`	1.71 GB
`UD-IQ4_XS`	1.74 GB
`Q4_K_S`	1.83 GB
`UD-IQ4_NL`	1.83 GB
`Q4_K_M`	1.93 GB
`UD-Q4_K_XL`	1.93 GB
`Q5_K_S`	2.17 GB
`UD-Q5_K_S`	2.17 GB
`Q5_K_M`	2.22 GB
`UD-Q5_K_M`	2.22 GB
`UD-Q5_K_XL`	2.22 GB
`Q6_K`	2.54 GB
`UD-Q6_K`	2.54 GB
`UD-Q6_K_XL`	2.54 GB
`Q8_0`	3.29 GB
`UD-Q8_K_XL`	3.29 GB
`F16`	6.18 GB

For a 3B‑parameter model, even the larger files are quite manageable. Here’s what I recommend: F16 (6.18 GB) or Q8_0 (3.29 GB).

The other quants are also usable!

Usage

With `ollama`

ollama run hf.co/constructai/VibeThinker-3B-GGUF:F16

With `llama.cpp`

llama-server -hf constructai/VibeThinker-3B-GGUF:VibeThinker-3B-GGUF-F16.gguf

or

llama-cli -hf constructai/VibeThinker-3B-GGUF:VibeThinker-3B-GGUF-F16.gguf

With `LM Studio`

lms get constructai/VibeThinker-3B-GGUF@F16

Downloads last month: 1,666

GGUF

Model size

3B params

Architecture

qwen2

Hardware compatibility

1-bit

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Model tree for constructai/VibeThinker-3B-GGUF

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-Coder-3B

Finetuned

WeiboAI/VibeThinker-3B

Quantized

(46)

this model