Downloads · 30 days
30
3% of all-time downloads
Qwen-3.5-28B-A3B-REAP-4bit-mlx
Alanfwilliams/Qwen-3.5-28B-A3B-REAP-4bit-mlx
Qwen-3.5-28B-A3B-REAP-4bit-mlx is a machine learning model from Alanfwilliams. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for mlx. The card lists the license as apache-2.0.
MLX 4-bit quantized version of 0xSero/Qwen-3.5-28B-A3B-REAP for Apple Silicon.
All-time downloads
877
Public
Parameters
28.2B
15.9 GB on disk
Likes
2
Public
Repo makeup
Click a slice to open those files.
.safetensors15.9 GB · 100%
Parameter types
How the weights are stored.
U3228.2B · 100%
Base models
Model card
From the Hugging Face model README
Qwen-3.5-28B-A3B-REAP-4bit (MLX)
MLX 4-bit quantized version of 0xSero/Qwen-3.5-28B-A3B-REAP for Apple Silicon.
Model Details
- Base model: Qwen3.5-35B-A3B with REAP expert pruning (20% pruned, 256 -> 205 experts)
- Quantization: 4-bit affine, group size 64
- Size on disk: ~16 GB
- Peak memory: ~16 GB
- Generation speed: ~64 tokens/sec (Apple M-series)
Usage
pip install mlx-lm
mlx_lm.generate --model Alanfwilliams/Qwen-3.5-28B-A3B-REAP-4bit-mlx --prompt "Hello"
Or serve as an OpenAI-compatible API:
mlx_lm.server --model Alanfwilliams/Qwen-3.5-28B-A3B-REAP-4bit-mlx --port 8080
Conversion
Converted using a custom script that handles REAP's pruned expert layout. The standard mlx_lm.convert doesn't support REAP models because the per-expert weight keys need to be fused into MLX's SwitchGLU format with the correct (pruned) expert count.
Credits
- Original model: 0xSero/Qwen-3.5-28B-A3B-REAP
- REAP pruning method: arXiv:2510.13999
- Base architecture: Qwen3.5-35B-A3B by Alibaba
- Library
- mlx
- Type
- qwen3_5_moe
- License
- apache-2.0
- Languages
- en
- Created
- Mar 31, 2026
- Updated
- Apr 1, 2026