Instructions to use poisonxa/Qwable-27B-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 poisonxa/Qwable-27B-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 poisonxa/Qwable-27B-GGUF:MXFP4
# Run inference directly in the terminal:
llama cli -hf poisonxa/Qwable-27B-GGUF:MXFP4

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf poisonxa/Qwable-27B-GGUF:MXFP4
# Run inference directly in the terminal:
llama cli -hf poisonxa/Qwable-27B-GGUF:MXFP4

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 poisonxa/Qwable-27B-GGUF:MXFP4
# Run inference directly in the terminal:
./llama-cli -hf poisonxa/Qwable-27B-GGUF:MXFP4

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 poisonxa/Qwable-27B-GGUF:MXFP4
# Run inference directly in the terminal:
./build/bin/llama-cli -hf poisonxa/Qwable-27B-GGUF:MXFP4

Use Docker

docker model run hf.co/poisonxa/Qwable-27B-GGUF:MXFP4

LM Studio
Jan
Ollama
How to use poisonxa/Qwable-27B-GGUF with Ollama:
```
ollama run hf.co/poisonxa/Qwable-27B-GGUF:MXFP4
```
Unsloth Desktop

How to use poisonxa/Qwable-27B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf poisonxa/Qwable-27B-GGUF:MXFP4

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": "poisonxa/Qwable-27B-GGUF:MXFP4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use poisonxa/Qwable-27B-GGUF with Docker Model Runner:
```
docker model run hf.co/poisonxa/Qwable-27B-GGUF:MXFP4
```

Lemonade

How to use poisonxa/Qwable-27B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull poisonxa/Qwable-27B-GGUF:MXFP4

Run and chat with the model

lemonade run user.Qwable-27B-GGUF-MXFP4

List all available models

lemonade list

Hermes Agent

How to use poisonxa/Qwable-27B-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 poisonxa/Qwable-27B-GGUF:MXFP4

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 poisonxa/Qwable-27B-GGUF:MXFP4

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use poisonxa/Qwable-27B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf poisonxa/Qwable-27B-GGUF:MXFP4

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 "poisonxa/Qwable-27B-GGUF:MXFP4" \
  --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"

Qwable-27B GGUF (benchmark files)

The two files used for the PXA cross-engine speed comparisons on Tesla P100 and V100 cards, published so every number in the pxq_llama release notes and bench/fair-battle.md can be reproduced from a download. Both are conversions of the same 27B dense hybrid (Qwen3.8-27B lineage, abliterated), so the engines are compared on identical weights.

File	Format	Size	sha256 (first 16)	Runs on
`Qwable-27B-PXQ4core.gguf`	PXQ4 (pxq_llama codec)	15.7 GB	`6fb4a437684e8bbe`	pxq_llama only
`Qwable-27B-MXFP4-lite.gguf`	MXFP4	15.0 GB	`1d04b75817358cf7`	mainline llama.cpp, ik_llama.cpp, pxq_llama

Protocol: bench/fair/protocol.md in the engine repo (temperature 0, median of 7, sha-checked files, unique prompts). Full sha256 sums are in checksums.sha256 here and in bench/fair/weights/MANIFEST.sha256 in the repo.

Community: Discord — PXA Network

Downloads last month: 548

GGUF

Model size

27B params

Architecture

qwen35

Hardware compatibility

4-bit

View +1 variant

Inference Providers NEW

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