Instructions to use oolabs/sommerfugl-31b-v2-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use oolabs/sommerfugl-31b-v2-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 oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf oolabs/sommerfugl-31b-v2-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 oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf oolabs/sommerfugl-31b-v2-GGUF: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 oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M

Use Docker

docker model run hf.co/oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M

Llama new
LM Studio
Jan
Ollama
How to use oolabs/sommerfugl-31b-v2-GGUF with Ollama:
```
ollama run hf.co/oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M
```
Unsloth Desktop

How to use oolabs/sommerfugl-31b-v2-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf oolabs/sommerfugl-31b-v2-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": "oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use oolabs/sommerfugl-31b-v2-GGUF with Docker Model Runner:
```
docker model run hf.co/oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M
```

Lemonade

How to use oolabs/sommerfugl-31b-v2-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.sommerfugl-31b-v2-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use oolabs/sommerfugl-31b-v2-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 oolabs/sommerfugl-31b-v2-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 oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use oolabs/sommerfugl-31b-v2-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf oolabs/sommerfugl-31b-v2-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 "oolabs/sommerfugl-31b-v2-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"

Sommerfugl-31B-v2 — GGUF

GGUF builds of Sommerfugl-31B-v2, a Norwegian language model by oolabs.no, for local use with llama.cpp, Ollama and LM Studio.

Built with Gemma. Sommerfugl-31B-v2 is a finetune of google/gemma-4-31B-it, modified by oolabs.no. Use is governed by the Gemma Terms of Use.

Results, known limitations and intended use are on the main model card. The chat template is embedded in every file.

Thinking mode is not supported. Sommerfugl is trained for direct answers; the embedded chat template always runs in non-thinking mode, so no extra flags are needed.

Files

file	quant	size	notes
`sommerfugl-31b-v2-Q4_K_M.gguf`	Q4_K_M	~19 GB	default; fits a 24 GB GPU or a 32 GB Mac
`sommerfugl-31b-v2-Q5_K_M.gguf`	Q5_K_M	~22 GB	a little closer to full quality
`sommerfugl-31b-v2-Q8_0.gguf`	Q8_0	~33 GB	near full quality

Quantization changes outputs slightly; the benchmark numbers on the main card are for the full-precision weights.

Usage

Ollama

ollama run hf.co/oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M

llama.cpp

llama-server -hf oolabs/sommerfugl-31b-v2-GGUF:Q4_K_M --jinja -ngl 99 -c 8192

LM Studio: search for oolabs/sommerfugl-31b-v2-GGUF and pick a quant.

Downloads last month: 66

GGUF

Model size

31B params

Architecture

gemma4

Hardware compatibility

4-bit

5-bit

8-bit

Inference Providers NEW

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

Model tree for oolabs/sommerfugl-31b-v2-GGUF

Base model

google/gemma-4-31B

Finetuned

google/gemma-4-31B-it

Finetuned

oolabs/sommerfugl-31b-v2

Quantized

(3)

this model