4-bit Open on Hugging Face
Parameters
1.3B
Weights (est.)
≈ 2.2 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/MiniCPM-V-4.6-nvfp4 --image photo.jpg --prompt "Describe this image"
Details
- Base model
- openbmb/MiniCPM-V-4.6 (quantized)
- Vendor
- openbmb
- Task
- Vision + text
- Quantisation
- 4-bit
- Licence
- apache-2.0
- Downloads
- 28 in the last 30 days · 509 all time · 0 likes
- Published
- 2026-05-11 · updated 2026-05-13 (6mo ago)
- Tags
- minicpmv4_6, minicpm-v, multimodal, On-Device Model
Other MLX builds of openbmb/MiniCPM-V-4.6
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| MiniCPM-V-4.6-4bit | 4-bit | ≈ 2.2 GB | 8 GB+ | 434 |
| MiniCPM-V-4.6-nvfp4 (this one) | 4-bit | ≈ 2.2 GB | 8 GB+ | 28 |
| MiniCPM-V-4.6-mxfp4 | MXFP4 | ≈ 2.2 GB | 8 GB+ | 28 |
| MiniCPM-V-4.6-5bit | 5-bit | ≈ 2.2 GB | 8 GB+ | 45 |
| MiniCPM-V-4.6-8bit | 8-bit | ≈ 2.3 GB | 8 GB+ | 68 |
| MiniCPM-V-4.6-mxfp8 | MXFP8 | ≈ 2.3 GB | 8 GB+ | 57 |
| MiniCPM-V-4.6-bf16 | bf16 | ≈ 2.6 GB | 8 GB+ | 196 |