Instructions to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF",
	filename="gemma-4-26B-A4B-it-qat-mxfp4.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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
# Run inference directly in the terminal:
llama-cli -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
# Run inference directly in the terminal:
llama-cli -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
# Run inference directly in the terminal:
./llama-cli -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
# Run inference directly in the terminal:
./build/bin/llama-cli -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

Use Docker

docker model run hf.co/FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

LM Studio
Jan
Ollama
How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with Ollama:
```
ollama run hf.co/FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
```

Unsloth Studio

How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF to start chatting

How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

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": "FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

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 FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with Docker Model Runner:
```
docker model run hf.co/FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE
```

Lemonade

How to use FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF:MXFP4_MOE

Run and chat with the model

lemonade run user.Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF-MXFP4_MOE

List all available models

lemonade list

Gemma 4 26B-A4B Instruct - QAT + MXFP4 Hybrid GGUF

QAT-optimized weights preserved at Q4_0, overhead tensors quantized to MXFP4.

What Makes This Different

This is a hybrid quantization of Google official QAT (Quantization-Aware Training) model. Instead of requantizing the Q4_0 weights (which breaks QAT benefits and vision quality), we:

Kept all weight tensors at Q4_0 - attention, FFN, embeddings - exactly as Google trained them
Quantized only the F32 norm/bias tensors to MXFP4 - these are the overhead tensors (layer norms, RMS norms, etc.)
Used Google QAT mmproj - the vision projector trained alongside the QAT model

Why Standard MXFP4 from QAT Breaks Vision

Google QAT model was specifically trained to be resilient to Q4_0 quantization patterns. The weight values learned during QAT compensate for Q4_0 rounding. When you requantize Q4_0 -> F32 -> MXFP4, a second round of quantization error is introduced that QAT training did not account for. Vision tokens flow through the same attention/FFN layers - precision loss disproportionately degrades vision.

How the Hybrid Approach Works

Using llama-quantize --tensor-type-file with --allow-requantize:

llama-quantize --allow-requantize --tensor-type-file keep_q4.txt input.gguf output.gguf MXFP4

The tensor-type-file lists all Q4_0/Q4_K tensors to keep at their current type. When the quantizer sees cur_type == new_type, it copies the tensor data as-is - zero precision loss. Only the remaining F32 tensors are quantized to MXFP4.

Usage

# llama.cpp
llama-server -m gemma-4-26B-A4B-it-qat-mxfp4.gguf --mmproj mmproj-gemma-4-26B-A4B-it-qat.gguf -ngl 99

Source

Base model: google/gemma-4-26B-A4B-it-qat-q4_0-unquantized
Quantized with: llama.cpp build 537 (commit d2c6795)
Chat template: Native Gemma 4 (thinking enabled by default)
Vision: Full multimodal support via QAT mmproj

Files

File	Description
gemma-4-26B-A4B-it-qat-mxfp4.gguf	Q4_0 weights + MXFP4 norms
mmproj-gemma-4-26B-A4B-it-qat.gguf	QAT vision projector

License

Apache 2.0 (same as base model)

Downloads last month: 732

GGUF

Model size

25B params

Architecture

gemma4

Hardware compatibility

4-bit

View +1 variant

Inference Providers NEW

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

Model tree for FreedomAISVR/Gemma-4-26B-A4B-it-QAT-MXFP4-GGUF

Base model

google/gemma-4-26B-A4B

Finetuned

google/gemma-4-26B-A4B-it

Finetuned

google/gemma-4-26B-A4B-it-qat-q4_0-unquantized

Quantized

(22)

this model