Instructions to use Wolkyen/Neptun3-One with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Wolkyen/Neptun3-One 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 Wolkyen/Neptun3-One:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Wolkyen/Neptun3-One:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Wolkyen/Neptun3-One:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Wolkyen/Neptun3-One: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 Wolkyen/Neptun3-One:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf Wolkyen/Neptun3-One:Q4_K_M

Use Docker

docker model run hf.co/Wolkyen/Neptun3-One:Q4_K_M

Llama new
LM Studio
Jan
Ollama
How to use Wolkyen/Neptun3-One with Ollama:
```
ollama run hf.co/Wolkyen/Neptun3-One:Q4_K_M
```
Unsloth Desktop

How to use Wolkyen/Neptun3-One with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Wolkyen/Neptun3-One: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": "Wolkyen/Neptun3-One:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Wolkyen/Neptun3-One with Docker Model Runner:
```
docker model run hf.co/Wolkyen/Neptun3-One:Q4_K_M
```

Lemonade

How to use Wolkyen/Neptun3-One with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Wolkyen/Neptun3-One:Q4_K_M

Run and chat with the model

lemonade run user.Neptun3-One-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use Wolkyen/Neptun3-One with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Wolkyen/Neptun3-One: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 Wolkyen/Neptun3-One:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Wolkyen/Neptun3-One with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Wolkyen/Neptun3-One: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 "Wolkyen/Neptun3-One: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"

Neptun3 One

Neptun3 One is a lightweight, 350-million-parameter language model developed by Wolkyen, designed to deliver strong general-purpose performance in an exceptionally compact architecture.

Built using a hybrid architecture, Neptun3 One focuses on balancing intelligence, efficiency, and practical usability. It is designed to compete with and outperform similarly sized language models across everyday tasks while maintaining a small computational footprint.

Despite its compact size, Neptun3 One supports native tool use, enabling applications to integrate external tools and build lightweight AI-powered workflows.

Key Features

350M Parameters: An exceptionally compact model designed for efficient inference.
Hybrid Architecture: Combines architectural techniques to balance performance and computational efficiency.
General-Purpose Intelligence: Designed for conversations, instruction following, text generation, and everyday assistance.
Lightweight Mathematics: Capable of handling basic mathematical reasoning and calculations.
Native Tool Use: Built-in support for tool calling and tool-integrated applications.
Efficient Deployment: Suitable for resource-constrained environments and lightweight AI applications.

Intended Use

Neptun3 One is designed for:

General-purpose AI assistants
Lightweight mathematical tasks
Tool-calling applications and simple agents
Text generation and instruction following
Local AI deployments and experimental applications

Performance

Neptun3 One is developed with the goal of outperforming models in its parameter class while maintaining efficient inference requirements.

Its hybrid architecture is designed to deliver a strong balance between model capability, speed, and resource usage.

Limitations

As a compact 350M-parameter model, Neptun3 One may struggle with complex reasoning, advanced mathematics, long multi-step tasks, and highly specialized knowledge.

Outputs should be verified when accuracy is important.

Developed by Wolkyen

Wolkyen develops fast, efficient, and accessible artificial intelligence models, with a focus on making capable AI available across a wide range of hardware.

What can you do with 10× the Speed?

Website: https://wolkyen.grok.me

Downloads last month: -

GGUF

Model size

0.4B params

Architecture

lfm2

Hardware compatibility

4-bit

6-bit

8-bit

16-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support