MXFP4 Open on Hugging Face
Parameters
27B
Weights (est.)
≈ 16.1 GB
Mac memory (est.)
24 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/Qwen3.6-27B-mxfp4 --image photo.jpg --prompt "Describe this image"
Details
- Base model
- Qwen/Qwen3.6-27B (quantized) · Qwen3.6 27B on the radar →
- Vendor
- Qwen
- Task
- Vision + text
- Quantisation
- MXFP4
- Licence
- apache-2.0
- Downloads
- 97 in the last 30 days · 10K all time · 5 likes
- Published
- 2026-04-22 · updated 2026-04-22 (6mo ago)
- Tags
- qwen3_5
Other MLX builds of Qwen/Qwen3.6-27B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen3.6-27B-MTP-4bit | 4-bit | ≈ 239 MB | 8 GB+ | 208 |
| Qwen3.6-27B-MTP-5bit | 5-bit | ≈ 292 MB | 8 GB+ | 49 |
| Qwen3.6-27B-MTP-bf16 | bf16 | ≈ 849 MB | 8 GB+ | 58 |
| Huihui-Qwen3.6-27B-abliterated-4.5bit-msq | 5-bit | ≈ 4.6 GB | 8 GB+ | 44 |
| Qwen3.6-27B-4bit | 4-bit | ≈ 16.1 GB | 24 GB+ | 49K |
| Qwen3.6-27B-OptiQ-4bit | 4-bit | ≈ 16.1 GB | 24 GB+ | 2.2K |
| Qwen3.6-27B-nvfp4 | 4-bit | ≈ 16.1 GB | 24 GB+ | 1.5K |
| Qwen3.6-27B-mxfp4 (this one) | MXFP4 | ≈ 16.1 GB | 24 GB+ | 97 |
| Qwen3.6-27B-5bit | 5-bit | ≈ 19.4 GB | 24 GB+ | 100 |
| Qwen3.6-27B-6bit | 6-bit | ≈ 22.8 GB | 32 GB+ | 213 |
| Qwen3.6-27B-8bit | 8-bit | ≈ 29.5 GB | 48 GB+ | 2.7K |
| Qwen3.6-27B-mxfp8 | MXFP8 | ≈ 29.5 GB | 48 GB+ | 749 |
| Qwen3.6-27B-bf16 | bf16 | ≈ 54.7 GB | 64 GB+ | 1.1K |