Gemma 4 12B IT is a multimodal model built by Google DeepMind that handles text, image, and audio inputs and generates text output. It is designed for efficient on-device and server deployment.

This repository contains GGUF conversions of the QAT-optimized checkpoint, making it usable with llama.cpp and other GGUF-compatible inference engines.

Model Architecture

Property	Value
Architecture	Gemma4ForConditionalGeneration
Parameters	12B
Layers	48
Embedding Dimension	3840
Feed Forward Length	15360
Attention Heads	16
Sliding Window	1024 tokens
Context Length	256K tokens (262144)
Vocabulary Size	262K (262144)
Supported Modalities	Text, Image, Audio
Attention	Hybrid (sliding window + global, every 6th layer)
RoPE	Proportional RoPE (p-RoPE) on global layers
Logit Softcapping	30.0

GGUF Files

File	Format	Size	Description
`gemma-4-12b-it-qat-Q4_0.gguf`	Q4_0	6.5G	QAT Q4_0 — native QAT quantization
`gemma-4-12b-it-qat-Q4_K_M.gguf`	Q4_K_M	6.9G	K-quant, medium
`gemma-4-12b-it-qat-Q4_K_S.gguf`	Q4_K_S	6.6G	K-quant, small
`gemma-4-12b-it-qat-Q5_K_M.gguf`	Q5_K_M	8.0G	K-quant, medium
`gemma-4-12b-it-qat-Q5_K_S.gguf`	Q5_K_S	7.8G	K-quant, small
`gemma-4-12b-it-qat-Q6_K.gguf`	Q6_K	9.2G	K-quant, higher precision
`gemma-4-12b-it-qat-Q8_0.gguf`	Q8_0	12G	8-bit, highest GGUF precision
`gemma-4-12b-it-qat-bf16.gguf`	bf16	23G	Full bfloat16 (unquantized)
`mmproj.gguf`	—	168M	Multimodal projector (vision + audio)

A chat_template.jinja file is also provided for use with chat-based inference.

Note on QAT: The Q4_0 file is the native QAT quantization. The K-quant and Q8_0 variants are additional GGUF quantizations produced from the QAT checkpoint. The QAT optimization preserves quality close to bfloat16 while dramatically reducing memory requirements.

Note on mmproj: The mmproj.gguf file contains the vision and audio projectors needed for multimodal (image/audio) inference. It is shared across all quantization variants.

Usage

llama.cpp (CLI)

bash

# Run text-only inference
./llama-cli \
  -m gemma-4-12b-it-qat-Q4_0.gguf \
  -p "Explain quantum computing in simple terms." \
  --temp 1.0 --top-k 64 --top-p 0.95

llama-server (OpenAI-compatible API)

bash

# Text-only
./llama-server \
  -m gemma-4-12b-it-qat-Q4_0.gguf \
  --host 0.0.0.0 --port 8080

# Multimodal (image + audio)
./llama-server \
  -m gemma-4-12b-it-qat-Q4_0.gguf \
  --mmproj mmproj.gguf \
  --host 0.0.0.0 --port 8080

Multimodal (Image / Audio)

For image and audio inputs, use llama-server or llama-cli with the --mmproj flag pointing to mmproj.gguf. Refer to your inference engine's documentation for passing image/audio data alongside text prompts.

Modality order tip: For best results, place image content before text and audio content after text in your prompt.

Generation Parameters

Recommended parameters from the model's generation_config.json:

Parameter	Value
Temperature	1.0
Top-K	64
Top-P	0.95
BOS Token ID	2
EOS Token ID	1
Pad Token ID	0
Mask Token ID	4

Thinking Mode

Gemma 4 supports configurable thinking (reasoning) mode:

Enable: Include the <|think|> token at the start of the system prompt.
Output format: When thinking is enabled, the model outputs internal reasoning followed by the final answer:
```
<|channel>thought
[Internal reasoning]
<channel|>
[Final answer]
```
Disable: Omit the <|think|> token.

Many libraries like Transformers and llama.cpp handle the chat template complexities automatically.

Key Features

Multimodal: Text, image, and audio understanding
Long Context: 256K token context window
Function Calling: Native support for structured tool use (agentic workflows)
Multilingual: Support for 140+ languages
Native System Prompt: Supports the system role for structured conversations

Acknowledgements

Original model: google/gemma-4-12B-it
QAT checkpoint: google/gemma-4-12B-it-qat-q4-0
Technical report: Gemma 4 Technical Report (arXiv:2607.02770)

Citation

bibtex

@misc{gemmateam2026gemma4,
      title={Gemma 4 Technical Report},
      author={Gemma Team},
      year={2026},
      eprint={2607.02770},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2607.02770},
}

License

Apache License 2.0

WhiskyAKM/Gemma-4-12B-it-qat-GGUF

Gemma 4 12B IT QAT — GGUF

Model Overview