4-bit Open on Hugging Face
Parameters
1.0B
Weights (est.)
≈ 1.2 GB
Mac memory (est.)
8 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-vlm on an Apple silicon Mac:
pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/LightOnOCR-3-1B-mlx-4bit --image photo.jpg --prompt "Describe this image"
Details
- Base model
- lightonai/LightOnOCR-3-1B (quantized)
- Vendor
- lightonai
- Task
- Vision + text
- Quantisation
- 4-bit
- Licence
- apache-2.0
- Downloads
- 189 in the last 30 days · 189 all time · 0 likes
- Published
- 2026-10-17 · updated 2026-10-17 (25d ago)
- Tags
- mistral3, document-understanding, visual-grounding, vision-language, tables, forms
Other MLX builds of lightonai/LightOnOCR-3-1B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| LightOnOCR-3-1B-mlx-4bit (this one) | 4-bit | ≈ 1.2 GB | 8 GB+ | 189 |
| LightOnOCR-3-1B-mlx-8bit | 8-bit | ≈ 1.5 GB | 8 GB+ | 248 |