Instructions to use poisonxa/PXA-Fusion4-35B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use poisonxa/PXA-Fusion4-35B-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="poisonxa/PXA-Fusion4-35B-GGUF",
	filename="PXA-Fusion4-35B-PXQ2.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use poisonxa/PXA-Fusion4-35B-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/PXA-Fusion4-35B-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf poisonxa/PXA-Fusion4-35B-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf poisonxa/PXA-Fusion4-35B-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf poisonxa/PXA-Fusion4-35B-GGUF:F16

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/PXA-Fusion4-35B-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf poisonxa/PXA-Fusion4-35B-GGUF:F16

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/PXA-Fusion4-35B-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf poisonxa/PXA-Fusion4-35B-GGUF:F16

Use Docker

docker model run hf.co/poisonxa/PXA-Fusion4-35B-GGUF:F16

LM Studio
Jan
Ollama
How to use poisonxa/PXA-Fusion4-35B-GGUF with Ollama:
```
ollama run hf.co/poisonxa/PXA-Fusion4-35B-GGUF:F16
```

Unsloth Studio

How to use poisonxa/PXA-Fusion4-35B-GGUF 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 poisonxa/PXA-Fusion4-35B-GGUF 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 poisonxa/PXA-Fusion4-35B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for poisonxa/PXA-Fusion4-35B-GGUF to start chatting

How to use poisonxa/PXA-Fusion4-35B-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/PXA-Fusion4-35B-GGUF:F16

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": "poisonxa/PXA-Fusion4-35B-GGUF:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use poisonxa/PXA-Fusion4-35B-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/PXA-Fusion4-35B-GGUF:F16

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/PXA-Fusion4-35B-GGUF:F16

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use poisonxa/PXA-Fusion4-35B-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/PXA-Fusion4-35B-GGUF:F16

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/PXA-Fusion4-35B-GGUF:F16" \
  --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 poisonxa/PXA-Fusion4-35B-GGUF with Docker Model Runner:
```
docker model run hf.co/poisonxa/PXA-Fusion4-35B-GGUF:F16
```

Lemonade

How to use poisonxa/PXA-Fusion4-35B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull poisonxa/PXA-Fusion4-35B-GGUF:F16

Run and chat with the model

lemonade run user.PXA-Fusion4-35B-GGUF-F16

List all available models

lemonade list

PXA-Fusion4-35B — 35B-A3B · abliterated · super-agentic · anti-fabrication

A 35B Mixture-of-Experts model tuned for agentic tool-use, coding, and uncensored reasoning — engineered to run fully GPU-resident on modest and salvaged hardware via PXA's PXQ quantization. Built on a Qwen3.5-35B-A3B foundation, refined through PXA's fusion + calibration pipeline. PXQ2–PXQ6 tiers fit anything from an 11 GB card up. Every quant ships with speculative-decode (MTP) heads and a vision projector.

Probe it. Pick at it. It doesn't fold — and it won't make things up.

Community: Discord — PXA Network · Engine: pxq_llama

Re-download if you pulled before 2026-07-24 - chat template fixed

The first upload shipped a chat template with an inverted enable_thinking default. Unless your client explicitly sent enable_thinking:false, the template pre-filled an unclosed <think> tag, so generation began inside an unbounded reasoning channel. In agentic and coding harnesses this looked like the model planning forever - long first-person prose ("I'll add X. I'll add Y."), no tool calls, no files written, eventually repeating itself until you killed it.

Fixed 2026-07-24. All quant tiers have been re-uploaded. Thinking is now opt-in, matching upstream Qwen: the default renders no-think, and enable_thinking:true still enables reasoning exactly as before.

The weights are unchanged. This was purely the embedded chat template - every tensor is byte-identical to the original upload, only template metadata changed. If you already had it working by passing enable_thinking:false, nothing changes for you.

Running an agent on long open-ended prompts? See Run it below - a thinking budget and Qwen's own presence-penalty recommendation matter a lot there.

What's new vs Fusion2

Fusion4 is the anti-fabrication generation. The headline gain is truthfulness under pressure without losing compliance — a differential graft that imported fabrication-resistance while keeping the model fully uncensored.

How it was built

Fusion4 is not a fine-tune with a new name. It's the output of a multi-stage model-engineering program — weight-space fusion, expert-level surgery, and adversarial-corpus hardening on a Qwen3.5-35B-A3B foundation.

Foundation — Qwen 3.5-35B-A3B MoE: 35B parameters, ~3B active per token, 256 experts × 40 layers, hybrid attention (small KV footprint).
Fusion + calibration — weight-space blend + imatrix calibration of the expert bank.
Anti-fabrication graft — a differential edit that raises false-premise rejection while leaving the uncensored behavior intact.
MTP grafting — a speculative-decode head on every shipped quant (inert unless --spec-type is passed).

Quants (pick by your VRAM)

Every quant is MTP-headed (spec-decode ready) and abliterated. Speeds are measured with the pxq_llama engine, PXA_ENHANCE=1, fair-battle protocol (cold ~5.8k-token prompt, temp 0, median of 3).

File	Tier	bpw	Size	Fits
`PXA-Fusion4-35B-PXQ6.gguf`	PXQ6 (real 5-bit)	~5.27	23.3 GB	24 GB card — top quality
`PXA-Fusion4-35B-PXQ4.gguf`	PXQ4	~4.27	19.3 GB	flagship — recommended, 2×16 GB or one 24 GB
`PXA-Fusion4-35B-PXQU16.gguf`	PXQ-Universal	~3.2	14.6 GB	fills a 16 GB card
`PXA-Fusion4-35B-PXQU12.gguf`	PXQ-Universal	~2.65	12.2 GB	fills a 12 GB card
`PXA-Fusion4-35B-PXQ2.gguf`	PXQ2	~2.27	11.2 GB	smallest — 11 GB cards
`mmproj-fusion4-f16.gguf`	vision projector	—	0.9 GB	add for image input

Measured speed (t/s), PXQ4:

Cards	Prefill	Decode	Decode + MTP
2× Tesla P100 (sm_60)	1,163	56.8	60.7 (accept ~0.68)
1× Tesla V100 (sm_70)	2,358	—	108.3
PXQ2, 1× P100	1,161	58.4	—
PXQ2, GTX 1080 Ti (ub 512)	855	61.6	—

Run it — exactly what we use

Get the pxq_llama engine (prebuilt release or build main), then:

llama-server -m PXA-Fusion4-35B-PXQ4.gguf \
  -c 73728 -ngl 99 -sm layer -fa on -ctk f16 -ctv f16 -b 2048 -ub 2048 \
  --jinja --chat-template-kwargs '{"enable_thinking":false}' \
  --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0
# env: PXA_ENHANCE=1 auto-tunes per card. Optional: --spec-type mtp:n_max=1 (MTP), --mmproj mmproj-fusion4-f16.gguf (vision)

The template flags are required, not optional. --jinja + --chat-template-kwargs '{"enable_thinking":false}' select the correct chat template and no-think mode. Omitting them is the #1 cause of looping / rambling / hallucination — this is a Qwen-family model and it needs its template.

Sampling we serve with: --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0 (tuned for multi-turn agent use). Running solo and want it more conservative? Use Qwen's official no-think set --temp 0.7 --top-p 0.8 --top-k 20 --min-p 0, and add --repeat-penalty 1.05 as a loop safety net.

Running an agentic / coding harness? Use these flags. They are not optional for agent work - they were derived from 144 interleaved measured runs, not guesswork.

llama-server -m PXA-Fusion4-35B-PXQ4.gguf --jinja \
  --reasoning off \
  --temp 0.7 --top-p 0.8 --top-k 20 --min-p 0 \
  --presence-penalty 1.5 --repeat-penalty 1.15 --repeat-last-n 256 --dry-multiplier 0.8 \
  -n 2048 -c 32768 -np 1 -ngl 99 -fa on

--reasoning off is the primary fix. The server default is --reasoning auto, which resolves to thinking ON, leaving every assistant turn starting inside an open <think> block with nothing to constrain it. Measured: 5/48 failures with thinking ON vs 0/72 with thinking OFF (Fisher p=0.0090).
The penalties are the brake. presence-penalty 1.5 is Qwen's own prescribed value for this model family; the other three are our PXQ1 repetition-guard preset, which does not auto-arm on other tiers. Verified not to break tool-calling (24/24 clean tool calls with penalties on).
-n 2048 is the belt. finish_reason currently reports "stop" even when generation is cut at the token cap, so a harness cannot detect a runaway. Cap it server-side.
Do NOT use --reasoning-budget 0. We measured it inert with this template - 24/24 runs still produced full-length reasoning. It will not save you.

Older build that rejects --reasoning? Use --chat-template-kwargs '{"enable_thinking":false}' instead.

Verify the flag actually applied before testing - the rendered prompt must end with a closed <think>\n\n</think>\n\n, not an open <think>\n:

curl -s http://127.0.0.1:8080/apply-template \
  -H "Content-Type: application/json" -d '{"messages":[{"role":"user","content":"hi"}]}'

Also make sure your client sends a real OpenAI tools array. Without one, the model emits its tool call as XML inside content, your harness sees an empty tool_calls, and it looks like the model did nothing. With the default reasoning_format=deepseek, prose goes to reasoning_content and content can be an empty string - read both.