---
language: en
library_name: mlx
tags:
- quantized
- mlx
base_model:
- Qwen/Qwen3.5-27B
pipeline_tag: text-generation
---
**See Qwen3.5-27B MLX in action - [demonstration video](https://youtu.be/OE5KdF4spss)**

#### Tested on a M3 Ultra 512GB RAM using [Inferencer app](https://inferencer.com)
- Single inference ~25.9 tokens/s @ 1000 tokens
- Vision inference: Not included in this language model (LM) only version
- Memory usage: ~22.1 GiB

*q7bit quant is expected to achieve higher than 96.96% token accuracy in our coding test*
| Quantization | Perplexity | Token Accuracy | Missed Divergence |
|:------------:|:----------:|:--------------:|:-----------------:|
| **q3.5**     | 168.0      | 43.45%         | 72.57%            |
| **q4.5**     | 1.33593    | 91.65%         | 27.61%            |
| **q5.5**     | 1.23437    | 95.05%         | 17.28%            |
| **q6.5**     | 1.21875    | 96.95%         | 12.03%            |
| **q8.5**     | 1.21093    | 97.55%         | 10.50%            |
| **q9**       | 1.21093    | 97.55%         | 10.50%            |
| **Base**     | 1.20312    | 100.0%         | 0.000%            |

- Perplexity: Measures the confidence for predicting base tokens (lower is better)
- Token Accuracy: The percentage of correctly generated base tokens
- Missed Divergence: Measures severity of misses; how much the token was missed by

##### Quantized with a modified version of [MLX](https://github.com/ml-explore/mlx)
##### For more details see [demonstration video](https://youtu.be/OE5KdF4spss) or visit [Qwen3.5-27B](https://huggingface.co/Qwen/Qwen3.5-27B).

## Disclaimer

We are not the creator, originator, or owner of any model listed. Each model is created and provided by third parties. Models may not always be accurate or contextually appropriate. You are responsible for verifying the information before making important decisions. We are not liable for any damages, losses, or issues arising from its use, including data loss or inaccuracies in AI-generated content.