Instructions to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="KikoCis/gemma-4-31b-it-IQ2_M-GGUF",
	filename="gemma4-31b-IQ2_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
# Run inference directly in the terminal:
llama-cli -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
# Run inference directly in the terminal:
llama-cli -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

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 KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
# Run inference directly in the terminal:
./llama-cli -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

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 KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

Use Docker

docker model run hf.co/KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

LM Studio
Jan

vLLM

How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "KikoCis/gemma-4-31b-it-IQ2_M-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KikoCis/gemma-4-31b-it-IQ2_M-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

Ollama
How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with Ollama:
```
ollama run hf.co/KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
```

Unsloth Studio new

How to use KikoCis/gemma-4-31b-it-IQ2_M-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 KikoCis/gemma-4-31b-it-IQ2_M-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 KikoCis/gemma-4-31b-it-IQ2_M-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for KikoCis/gemma-4-31b-it-IQ2_M-GGUF to start chatting

Pi new

How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

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": "KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use KikoCis/gemma-4-31b-it-IQ2_M-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 KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

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 KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

Run Hermes

hermes

Docker Model Runner
How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with Docker Model Runner:
```
docker model run hf.co/KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M
```

Lemonade

How to use KikoCis/gemma-4-31b-it-IQ2_M-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull KikoCis/gemma-4-31b-it-IQ2_M-GGUF:IQ2_M

Run and chat with the model

lemonade run user.gemma-4-31b-it-IQ2_M-GGUF-IQ2_M

List all available models

lemonade list

Gemma 4 31B IT — IQ2_M (CLI-focused imatrix)

Custom IQ2_M quantization of google/gemma-4-31b-it with an importance matrix calibrated on CLI / shell-assistant traces.

Size: 10.17 GB (6× smaller than f16)
BPW: 2.84
Layers: 60 (unmodified)
Calibration: imatrix from 4 chunks of CLI/bash-focused text
Tool: llama.cpp (llama-quantize --imatrix ...)

NL2Bash benchmark (50 examples, Stanford/Tellina test split)

Metric	This IQ2_M	f16 baseline	Unsloth UD-IQ2_M
Size	10.17 GB	61.4 GB	10.75 GB
Char-F1	84.71%	84.76%	84.02%
BLEU-1	44.72	43.94	42.38
BLEU-2	34.36	33.84	31.73
BLEU-4	22.39	21.02	18.64
Exact	12%	12%	10%

This IQ2_M matches f16 Char-F1 and beats f16 on BLEU-4 (22.39 vs 21.02) at 6× smaller. Also beats Unsloth UD-IQ2_M on every metric at 0.6 GB smaller.

Run with: llama-cli -m gemma4-31b-IQ2_M.gguf -ngl 99 --ctx-size 4096 --temp 0.1

Cross-model benchmark comparison (first public eval)

Note on BFCL scores: All BFCL scores reported here use our internal simplified evaluation (single-function-call subset with custom prompt/scoring), NOT the official Berkeley Function Calling Leaderboard methodology. Our scores are not directly comparable to the official leaderboard. We are working on running the official BFCL evaluation for comparable numbers.

Model	Size	Active	HumanEval+	MBPP+	BFCL v3	NL2Bash F1
This repo (Gemma 4 31B IQ2_M)	10.4 GB	31B	88.41%	82.01%	92.25%	84.71%
Qwen3.6 Q8_0 (≈f16)	35.2 GB	3B	81.10%	82.80%	95.25%	—
Qwen3.6 IQ2_M (sibling)	11.1 GB	3B	80.49%	78.31%	94.75%	81.63%
Gemma 4 E4B Q8_0	7.8 GB	4.5B	73.78%	73.28%	93.75%	79.75%

This model is the best available for code generation in the 10-12 GB tier — HumanEval+ 88.41% beats every Qwen3.6 variant including full precision.

How this compares

Full comparison (8 quantizations, layer-importance study, ablation charts) at otter.utopiaia.com and github.com/KikoCisBot/gemma4-31b-study.

At ~2.7 BPW, standard Q2_K collapses (F1 58.6%). Adaptive IQ2_M with a CLI-tuned imatrix holds 84.7% Char-F1 — functionally equivalent to f16 for shell-command generation.

Quickstart

# Download
huggingface-cli download KikoCis/gemma-4-31b-it-IQ2_M-GGUF gemma4-31b-IQ2_M.gguf --local-dir .

# Run
llama-cli -m gemma4-31b-IQ2_M.gguf -ngl 99 --ctx-size 4096 \
  --temp 0.1 -p "List all processes using port 8080"

Files

gemma4-31b-IQ2_M.gguf — quantized weights
gemma31b-imatrix.dat — importance matrix used for calibration

Citation

If you find this useful, consider starring the otter repo.

Real-World Agent Test Warning (April 2026)

Benchmark scores do not predict agent capability. In Docker-based autonomous testing, fine-tuned E4B models (95% BFCL) scored 0/10 while the unfine-tuned base scored 6/10. Fine-tuning for BFCL destroyed general reasoning (error recovery, strategy adaptation, anti-repetition). Fine-tuned E4B models have been withdrawn.

For autonomous agent tasks, use the base Gemma 4 model or a larger model at higher BPW. See: The Benchmark Trap — Full Study

Downloads last month: 292

GGUF

Model size

31B params

Architecture

gemma4

Hardware compatibility

2-bit