Instructions to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic 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 FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
# Run inference directly in the terminal:
llama cli -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
# Run inference directly in the terminal:
llama cli -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

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 FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
# Run inference directly in the terminal:
./llama-cli -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

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 FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

Use Docker

docker model run hf.co/FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

LM Studio
Jan
Ollama
How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with Ollama:
```
ollama run hf.co/FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
```
Unsloth Desktop

How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

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": "FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with Docker Model Runner:
```
docker model run hf.co/FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0
```

Lemonade

How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

Run and chat with the model

lemonade run user.Qwen-3.5-4B-A90-R10-Heretic-Q2_0

List all available models

lemonade list

Hermes Agent

How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

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 FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0

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 "FenrirLupus/Qwen-3.5-4B-A90-R10-Heretic:Q2_0" \
  --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"

Qwen3.5-4B-Heretic-A90-R10

This model is a specialized fine-tune of Qwen/Qwen3.5-4B engineered using the Heretic Model framework. It features a finely tuned behavioral alignment profile optimized for specific operational constraints.

What is a Heretic Model?

A Heretic Model is designed to shift the traditional boundaries of safety alignment and task adherence. Instead of utilizing standard, rigid alignment filters that can cause unnecessary refusals on complex or edge-case prompts, this model relies on a calibrated ratio of target behaviors.

The core metrics of this model are defined by the A/R naming convention:

A (Acceptance): The model's rate of compliance and willingness to fulfill complex, unconventional, or nuanced user prompts.
R (Refusal): The model's rate of strict boundary enforcement and alignment-driven declines.

Performance Profile

Based on benchmark testing, this model achieves the following behavioral distribution:

90% Acceptance (A90): High baseline helpfulness and reduced false-positive refusal rates on edge-case prompts.
10% Refusal (R10): A minimized but present guardrail system to prevent severe harm or explicitly unsafe outputs.

Intended Use

This model is intended for researchers and developers exploring the boundaries of LLM alignment, steerability, and the trade-offs between helpfulness and harmlessness.