Instructions to use Prevolut/socratic-gemma-4-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Prevolut/socratic-gemma-4-it with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Prevolut/socratic-gemma-4-it",
	filename="socratic-gemma-4-2B-it-Q4_K_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 Prevolut/socratic-gemma-4-it with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Prevolut/socratic-gemma-4-it:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Prevolut/socratic-gemma-4-it:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Prevolut/socratic-gemma-4-it:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Prevolut/socratic-gemma-4-it:Q4_K_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 Prevolut/socratic-gemma-4-it:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Prevolut/socratic-gemma-4-it:Q4_K_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 Prevolut/socratic-gemma-4-it:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Prevolut/socratic-gemma-4-it:Q4_K_M

Use Docker

docker model run hf.co/Prevolut/socratic-gemma-4-it:Q4_K_M

LM Studio
Jan

vLLM

How to use Prevolut/socratic-gemma-4-it with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Prevolut/socratic-gemma-4-it:Q4_K_M

Ollama
How to use Prevolut/socratic-gemma-4-it with Ollama:
```
ollama run hf.co/Prevolut/socratic-gemma-4-it:Q4_K_M
```

Unsloth Studio new

How to use Prevolut/socratic-gemma-4-it 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 Prevolut/socratic-gemma-4-it 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 Prevolut/socratic-gemma-4-it to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Prevolut/socratic-gemma-4-it to start chatting

Pi new

How to use Prevolut/socratic-gemma-4-it with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Prevolut/socratic-gemma-4-it:Q4_K_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": "Prevolut/socratic-gemma-4-it:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Prevolut/socratic-gemma-4-it with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Prevolut/socratic-gemma-4-it:Q4_K_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 Prevolut/socratic-gemma-4-it:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use Prevolut/socratic-gemma-4-it with Docker Model Runner:
```
docker model run hf.co/Prevolut/socratic-gemma-4-it:Q4_K_M
```

Lemonade

How to use Prevolut/socratic-gemma-4-it with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Prevolut/socratic-gemma-4-it:Q4_K_M

Run and chat with the model

lemonade run user.socratic-gemma-4-it-Q4_K_M

List all available models

lemonade list

🧠 Socratic-Gemma-4-IT (E2B) - Prevolut Ltd

This is a highly optimized, fine-tuned version of Google's Gemma 4 E2B IT (Edge 2B), developed and trained by Prevolut Ltd.

We engineered this model to bridge the gap between lightweight edge-computing and advanced structural reasoning. By utilizing a socratic fine-tuning approach (including high-quality datasets like GSM8K), this model excels at deterministic formatting, logical sequence tracking, and flawless tool orchestration.

🎯 Key Features & Enhancements

Socratic Reasoning Engine: Instead of guessing answers, the model is trained to break down complex, multi-stage system problems step-by-step, running internal plausibility checks before outputting the final result.
Format & Syntax Discipline: Highly disciplined in maintaining strict output structures. It isolates mathematical formulas cleanly and is exceptionally stable at generating pure JSON blocks without conversational clutter.
MCP & Tool Orchestration Ready: Due to its strict formatting adherence, this model is an ideal candidate for serving as a local agent interacting with the Model Context Protocol (MCP), executing API calls, and managing local system states (e.g., Docker, databases).
Multilingual Capability: Fully capable of reasoning and conversing in English, German, and French.
Edge Optimized: Exported in the highly efficient Q4_K_M GGUF format, ensuring lightning-fast inference on local workstations, mobile environments, and consumer hardware.

💻 Intended Use Cases

Local AI Agents: Powering privacy-first, on-device assistants.
System Orchestration: Translating natural language into structured JSON payloads for tool execution.
Complex Logic Tasks: Solving riddles, dynamic queue simulations, and multi-variable logic puzzles.

🛠️ Technical Specifications

Base Model: google/gemma-4-E2B-it
Architecture: 2 Billion Parameters (Edge-optimized)
Format: GGUF (Q4_K_M quantization)
License: Apache 2.0 (Fully cleared for commercial use)

🚀 How to use

You can load this model directly into standard local inference tools such as LM Studio, Ollama, or any application built on top of llama.cpp.

Example Prompt for Tool Execution

To leverage the model's structural discipline for tool calls, we recommend enforcing markdown code blocks in your system prompts:

You are a local system agent. If you need to use a tool, output ONLY a valid JSON block inside markdown formatting. Do not add any conversational text before or after the JSON.

Developed with focus on local AI efficiency by Prevolut Ltd

Downloads last month: 48

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

4-bit

Model tree for Prevolut/socratic-gemma-4-it

Base model

google/gemma-4-E2B

Finetuned

google/gemma-4-E2B-it

Quantized

(200)

this model

Prevolut
/

socratic-gemma-4-it