Qwen3.8-27B-GGUF by ggml-org

Qwen3.8-27B-GGUF by ggml-org

Runs Qwen3.8-27B, a 26.9B model that takes images and text and replies in text, as a GGUF file for local use through the llama tools. Made for people who want it on their own computer.

Made by ggml-org 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 Qwen3.8-27B, a model that takes images and text and answers in text. ggml-org made it with an automatic conversion step. Hugging Face estimates the weights at about 13.4 GB at 4-bit, and the page itself holds little beyond the run steps.

Who it suits: It suits people who want to try a 26.9B image-and-text model on their own machine through the llama tools.

Title: Who uses this model
Industry | Who uses it | How they use it | Example
Software | Developers | Serve a local model that reads images and text | Start the serve command and send it a question
Research | Machine learning hobbyists | Test how a 26.9B model behaves in GGUF form | Compare it with other quantized copies
Education | Students | Learn how local model serving works | Run the serve command and chat with the model

If you need a documented model with benchmarks, you can skip it, because the page says little about the model.

Where to use it:

• The Llama app
• A terminal, with the llama serve command
• Windows, macOS or Linux with Python, for downloading the files

Title: Jobs it does
Use case | Who | What they get back
Local chat | Anyone with a capable computer | Text replies from a model running on their own machine
Image questions | Developers | Text answers about pictures they provide
File download | Tinkerers | The GGUF files saved to a local folder

How to set it up: You need a way to run the model and enough disk space for it.

• The Llama app, or the llama command in a terminal
• Python with pip, if you only want the files
• Room for roughly 13.4 GB of weights, which is an estimate

1. Get the Llama app or make sure the llama command works in your terminal.
2. Run the serve command with the repository name, and the model loads from it.
3. To only save the files, install huggingface_hub and run the download command instead.

llama serve -hf ggml-org/Qwen3.8-27B-GGUF

pip install -U huggingface_hub
hf download ggml-org/Qwen3.8-27B-GGUF

How to use it: Start the serve command and the model comes up from the repository it names. The Llama app does the same job without a terminal. The download command saves the files if you prefer to keep a local copy.

Title: Main commands
Tool | What it does
llama serve | Starts the model from the repository named in the command
hf download | Saves the repository's files to your computer

Example: You start the model with the serve command, then give it a screenshot and ask what it shows. As an image-and-text model, it is meant to answer in plain text.

For newcomers and experts: A model like this reads a picture plus a question and writes back an answer. The easiest first step is the Llama app, which runs it without any commands. If you prefer a terminal, the serve command does the same.

GGUF is a single-file format for running models locally. Storing weights at 4-bit trades some precision for a smaller download, and the 13.4 GB figure is Hugging Face's estimate. The base model is Qwen/Qwen3.8-27B, and the card also lists a second source model named Qwen3.8-27B-DFlash2 by incoai.

Qwen3.8-27B base model
Automatic GGUF conversion
ggml-org repository
llama serve or the Llama app
Local replies in text

Good to know: The card is short and carries a to-do note for more information, so it gives no benchmarks or usage limits. Four quantized versions by other people exist, and none are fine-tunes.

• License: Apache-2.0.
• Cost to run: none named.
• Data it reaches: the files come from Hugging Face, and the model then runs on your own computer.
• Not for: the card names no restrictions, so check the original Qwen model page before relying on it.

On Hugging Face, 13 November 2026

• Task: image text to text
• Parameters: 26.9B
• Weights: about 13.4 GB at 4-bit (estimate)
• Formats: GGUF
• Downloads in the last 30 days: 1,000,687
• Likes: 157
• License: apache-2.0
• Based on: Qwen/Qwen3.8-27B
• Quantized versions: 4
• Tags: gguf, quantized, image-text-to-text, conversational

Links

• Model: https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF
• Files: https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download ggml-org/Qwen3.8-27B-GGUF

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