gemma-3-4b-it-GGUF

gemma-3-4b-it-GGUF

Answers questions about text and images on your own computer, as a GGUF copy of Google's Gemma 3 4B instruction model packaged by Unsloth. For developers and users who want a small local model.

Made by unsloth on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: This is a GGUF copy of Google's Gemma 3 4B instruction-tuned model, packaged by Unsloth. Gemma 3 takes text and images in and writes text out. The 4B size has a 128K-token context, and the card lists support for more than 140 languages. The record puts the model at 3.9B parameters, around 1.9 GB at 4-bit as an estimate.

Who it suits: It suits developers and curious users who want a small multimodal model on a laptop or desktop instead of a hosted service.

Industry | Who uses it | How they use it | Example
Software | Developers | Add a small local assistant that reads text and images | Ask a question about a picture
Productivity | Everyday users | Condense documents on their own computer | Summarize a long report
Education | Students of ML | Try a compact open model on reasoning tasks | Run question answering locally

Anyone who needs audio input can skip it, since the card lists text and images as the inputs.

Where to use it:

• Tools that read the GGUF format
• Laptops, desktops or your own cloud servers
• Windows, macOS or Linux, where the files download with Python

Use case | Who | What they get back
Question answering | Developers | A written answer to your prompt
Summarizing | Everyday users | A shorter version of a document you provide
Image understanding | Developers | A text description or answer about an image

How to set it up:

• Python with pip, for the download
• A runtime that can load GGUF files
• Memory for the model, roughly 1.9 GB when quantized to 4 bits

1. Install the Hugging Face hub client.
2. Download the repository's files.
3. Open the GGUF file in your runtime.

pip install -U huggingface_hub
hf download unsloth/gemma-3-4b-it-GGUF

How to use it: Load the GGUF file in your runtime and send it text, images or both. Each image is resized to 896 x 896 and costs 256 tokens of context. The model answers in text.

Tool | What it does
Text input | A question, a task description or a file to condense
Image input | Resized to 896 x 896 and encoded to 256 tokens each
Text output | A generated answer, summary or description

Example: You load the file, attach a photo and ask what it shows. The model returns a short text answer about the picture.

For newcomers and experts: A GGUF file is a single packed model that local tools can open. Start by loading it and asking a plain question before you try an image.

The 4B and 12B and 27B sizes share a 128K-token input context, while the 1B size has 32K. Unsloth supplies the quantized files, and its notebooks cover fine-tuning Gemma 3 models, which is a separate step from running this file.

Text and images go in
Images shrink to 896 x 896
The model reads up to 128K tokens
Text comes out

Good to know:

• Context: 128K tokens for the 4B size
• Images: each one takes 256 tokens
• Languages: the card says more than 140, and the record lists English
• License: gemma, so Google's Gemma terms of use apply
• Cost to run: none named
• Data it reaches: the text and images you give it, with no web or account access named
• Not for: nothing named beyond the Gemma terms of use

On Hugging Face, 13 November 2026

• Task: image text to text
• Library: transformers
• Parameters: 3.9B
• Weights: about 1.9 GB at 4-bit (estimate)
• Formats: GGUF
• Downloads in the last 30 days: 160,720
• Likes: 216
• License: gemma
• Languages: en
• Based on: google/gemma-3-4b-it
• Quantized versions: 3
• Tags: transformers, gguf, gemma3, image-text-to-text, unsloth, gemma, google, conversational

Links

• Model: https://huggingface.co/unsloth/gemma-3-4b-it-GGUF
• Files: https://huggingface.co/unsloth/gemma-3-4b-it-GGUF/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download unsloth/gemma-3-4b-it-GGUF

Then follow the usage in its model card. Nothing in this kit runs by itself.