MLX Community · First MLX build 2 Mar 2026
Qwen3.5-2B
MLX
TextVision apache-2.0
A 2.2B parameter vision and text model, with 16 MLX builds, from 3-bit to bf16.
- Parameters
- 2.2B From Hugging Face
- Quantisation
- 6-bit Qwen3.5-2B-6bit
- Memory needed
- 5 GB Estimate, unified memory
- Download
- 2.2 GB Weights, estimated
- Downloads
- 12K This build, last 30 days
- Task
- Vision and text Text, Vision
- MLX builds
- 16 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 5 GB at 6-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/Qwen3.5-2B-6bit --image photo.jpg --prompt "Describe this image"MLX builds
16 on mlx-community| Qwen3.5-2B-3bit | 3-bit | 859 MB | 3 GB | 35 |
|---|---|---|---|---|
| Qwen3.5-2B-mxfp4 | MXFP4 | 862 MB | 3 GB | 25 |
| Qwen3.5-2B-mixed_3_4 | 4-bit | 892 MB | 3 GB | 36 |
| Qwen3.5-2B-mixed_2_6 | 4-bit | 894 MB | 3 GB | 21 |
| Qwen3.5-2B-mixed_3_5 | 4-bit | 904 MB | 4 GB | 20 |
| Qwen3.5-2B-mixed_3_6 | 4-bit | 916 MB | 4 GB | 14 |
| Qwen3.5-2B-nvfp4 | 4-bit | 928 MB | 4 GB | 1.5K |
| Qwen3.5-2B-mixed_4_6 | 4-bit | 937 MB | 4 GB | 25 |
| Qwen3.5-2B-mixed_3_8 | 4-bit | 939 MB | 4 GB | 30 |
| Qwen3.5-2B-mixed_4_8 | 4-bit | 960 MB | 4 GB | 27 |
| Qwen3.5-2B-mxfp8 | MXFP8 | 1.2 GB | 4 GB | 747 |
| Qwen3.5-2B-4bit | 4-bit | 1.7 GB | 4 GB | 8.6K |
| Qwen3.5-2B-5bit | 5-bit | 2.0 GB | 5 GB | 99 |
| Qwen3.5-2B-6bit | 6-bit | 2.2 GB | 5 GB | 12K |
| Qwen3.5-2B-8bit | 8-bit | 2.7 GB | 5 GB | 509 |
| Qwen3.5-2B-bf16 | bf16 | 4.4 GB | 7 GB | 1.1K |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.