Instructions to use Luigi/minicpm5-1b-cursor-experiments with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Luigi/minicpm5-1b-cursor-experiments 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 Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Luigi/minicpm5-1b-cursor-experiments:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Luigi/minicpm5-1b-cursor-experiments: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 Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Luigi/minicpm5-1b-cursor-experiments: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 Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Luigi/minicpm5-1b-cursor-experiments:Q4_K_M

Use Docker

docker model run hf.co/Luigi/minicpm5-1b-cursor-experiments:Q4_K_M

LM Studio
Jan
Ollama
How to use Luigi/minicpm5-1b-cursor-experiments with Ollama:
```
ollama run hf.co/Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
```

Unsloth Studio

How to use Luigi/minicpm5-1b-cursor-experiments 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 Luigi/minicpm5-1b-cursor-experiments 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 Luigi/minicpm5-1b-cursor-experiments to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Luigi/minicpm5-1b-cursor-experiments to start chatting

How to use Luigi/minicpm5-1b-cursor-experiments with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Luigi/minicpm5-1b-cursor-experiments: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": "Luigi/minicpm5-1b-cursor-experiments:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use Luigi/minicpm5-1b-cursor-experiments with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Luigi/minicpm5-1b-cursor-experiments: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 "Luigi/minicpm5-1b-cursor-experiments: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"

Docker Model Runner
How to use Luigi/minicpm5-1b-cursor-experiments with Docker Model Runner:
```
docker model run hf.co/Luigi/minicpm5-1b-cursor-experiments:Q4_K_M
```

Lemonade

How to use Luigi/minicpm5-1b-cursor-experiments with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Luigi/minicpm5-1b-cursor-experiments:Q4_K_M

Run and chat with the model

lemonade run user.minicpm5-1b-cursor-experiments-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use Luigi/minicpm5-1b-cursor-experiments with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Luigi/minicpm5-1b-cursor-experiments: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 Luigi/minicpm5-1b-cursor-experiments:Q4_K_M

Run Hermes

hermes

Atomic Chat

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

MiniCPM5-1B-CURSOR — experiment checkpoints (measured negatives / regressed passes)

These are ALL trained weights from the 2026-08 MiniCPM round that were NOT selected for release. They exist for reproducibility and for the measured-negative record. Do not deploy any of these — each entry states why it lost to the published artifacts (p13/p15d on Luigi/minicpm5-1b-cursor).

file	pass	why not shipped
minicpm5-1b-p11-e1.Q4_K_M.gguf	p11 epoch-1	SYNTH-boost data; zh trap broke by epoch 2; e1 kept as probe
minicpm5-1b-p11-final.Q4_K_M.gguf	p11 (checkpoint-284)	zh trap FAIL
minicpm5-1b-p12.Q4_K_M.gguf	p12	raw 4/20 (over-assertion negatives); p13 supersedes
minicpm5-1b-p14.Q4_K_M.gguf	p14	zh trap FAIL (en-heavy K2 dose)
minicpm5-1b-p14b.Q4_K_M.gguf	p14b	en chain FAIL + zh trap FAIL
minicpm5-1b-p15.Q4_K_M.gguf	p15	en chain FAIL + zh trap FAIL (coverage dose too heavy)
minicpm5-1b-p15b.Q4_K_M.gguf	p15b	en chain FAIL
minicpm5-1b-p15c.Q4_K_M.gguf	p15c	G1 PASS but real-meeting coverage washed out (ACTIONS empty)
minicpm5-1b-p15e.Q4_K_M.gguf	p15e	en chain FAIL (DECISIONS dose)
minicpm5-1b-multirole-e1.Q4_K_M.gguf	multi-role experiment	measured negative: critic role 38% agreement
minicpm5-1b-multirole-e2.Q4_K_M.gguf	multi-role experiment	measured negative (same)
verifier-gemma-270m.Q4_K_M.gguf	gemma-3-270m verifier	70% agreement ceiling — 270M too small
verifier-granite-gr2.Q4_K_M.gguf	granite in-stream-form variant	90% claim-mode (worse than the published 97%)
critic-r8.lora / critic-r32.lora	LoRA critic adapters (r8/r32)	64% / 85% — generation bias resists low-rank correction

The published pair: Luigi/minicpm5-1b-cursor (p15d, p13) + Luigi/granite-4.0-350m-verifier.

Downloads last month: 176

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

Inference Providers NEW

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