MLX Community · First MLX build 28 Jan 2026
ERNIE-4.5-VL-28B-A3B-Thinking
MLX
TextVision apache-2.0
A 29B parameter vision and text model, with 5 MLX builds, from 4-bit to bf16.
- Parameters
- 29B From Hugging Face
- Quantisation
- 4-bit ERNIE-4.5-VL-28B-A3B-Thinking-4bit
- Memory needed
- 22 GB Estimate, unified memory
- Download
- 17.4 GB Weights, estimated
- Downloads
- 64 This build, last 30 days
- Task
- Vision and text Text, Vision
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 22 GB at 4-bit
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · mlx-vlm$ pip install -U mlx-vlm
$ python -m mlx_vlm.generate --model mlx-community/ERNIE-4.5-VL-28B-A3B-Thinking-4bit --image photo.jpg --prompt "Describe this image"MLX builds
5 on mlx-community| ERNIE-4.5-VL-28B-A3B-Thinking-4bit | 4-bit | 17.4 GB | 22 GB | 64 |
|---|---|---|---|---|
| ERNIE-4.5-VL-28B-A3B-Thinking-5bit | 5-bit | 21.0 GB | 26 GB | 36 |
| ERNIE-4.5-VL-28B-A3B-Thinking-6bit | 6-bit | 24.6 GB | 30 GB | 35 |
| ERNIE-4.5-VL-28B-A3B-Thinking-8bit | 8-bit | 31.8 GB | 38 GB | 42 |
| ERNIE-4.5-VL-28B-A3B-Thinking-bf16 | bf16 | 58.8 GB | 68 GB | 36 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.