---
language: en
license: apache-2.0
license_link: https://huggingface.co/Qwen/Qwen3.6-27B/blob/main/LICENSE
pipeline_tag: text-generation
tags:
- mlx
library_name: mlx
base_model: Qwen/Qwen3.6-27B
---

This model was converted to MLX format from [Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B) using [oMLX v0.3.6](https://github.com/jundot/omlx/releases/tag/v0.3.8).

## Settings

- Level: oQ4
- Sensitivity model: Qwen3.6-27B-MLX-Q8
- Text Only: yes
- Non-quant weight dtype: bfloat16

## What is "oQ"?

See ["oQ: oMLX Universal Dynamic Quantization"](https://github.com/jundot/omlx/blob/main/docs/oQ_Quantization.md) for details.

## I have M1/M2 Apple Silicon

Consider using specially optimized [deepsweet/Qwen3.6-27B-MLX-oQ4-FP16](https://huggingface.co/deepsweet/Qwen3.6-27B-MLX-oQ4-FP16) version with a very noticeable prompt processing boost. See [jundot/omlx/issues/604](https://github.com/jundot/omlx/issues/604) for details.

Otherwise stick with this one if you have M3+ Apple Silicon.

## I need a multimodal variant with Vision-Language

Feel free to request deepsweet/Qwen3.6-27B-MLX-VL-oQ4.