Instructions to use mlx-community/Mage-VL-8bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/Mage-VL-8bit with MLX:
# Make sure mlx-vlm is installed # pip install --upgrade mlx-vlm from mlx_vlm import load, generate from mlx_vlm.prompt_utils import apply_chat_template from mlx_vlm.utils import load_config # Load the model model, processor = load("mlx-community/Mage-VL-8bit") config = load_config("mlx-community/Mage-VL-8bit") # Prepare input image = ["http://images.cocodataset.org/val2017/000000039769.jpg"] prompt = "Describe this image." # Apply chat template formatted_prompt = apply_chat_template( processor, config, prompt, num_images=1 ) # Generate output output = generate(model, processor, formatted_prompt, image) print(output) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
mac mini / vMLX error
create session
Process exited before becoming ready: ValueError: Model type mage_vl not supported. Error: No module named 'mlx_vlm.speculative.drafters.mage_vl'
Is the error reported when using vMLX to load the model due to vMLX not supporting it?
Not vMLX β your mlx-vlm is just older than the model.
mage_vl was merged into mlx-vlm on 2026-07-29 (PR #1745), and the newest PyPI release (0.6.8) went out 2026-07-27 β two days earlier. So pip install mlx-vlm doesn't include it yet. Until 0.6.9 ships:
pip install -U git+https://github.com/Blaizzy/mlx-vlm
The speculative.drafters part of the message is a red herring. The loader tries mlx_vlm.models.<type> first and mlx_vlm.speculative.drafters.<type> second, then reports only the last error β so the failure that actually matters is the first one, mlx_vlm.models.mage_vl not being there. Mage-VL has no speculative drafter and doesn't need one.
One caveat: if vMLX bundles or pins its own copy of mlx-vlm rather than using your environment's, updating it there is what counts β or wait for 0.6.9 and a vMLX bump.