---
title: Qwen3.5-0.8B-LM-MLX-9bit
canonical_url: "https://www.modelscope.cn/models/inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit"
md_url: "https://www.modelscope.cn/models/inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit.md"
repository: inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit
last_updated: 2026-07-18
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3_5
architectures:
  - Qwen3_5ForConditionalGeneration
base_model:
  - Qwen/Qwen3.5-0.8B
base_model_relation: quantized
parameters: 235.5M
tensor_type:
  - BF16
  - U32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 126
stars: 0
tags:
  - quantized
  - mlx
---

# Qwen3.5-0.8B-LM-MLX-9bit

> Qwen3.5-0.8B-LM-MLX-9bit - inferencerlabs 在 ModelScope 开源的模型。See Qwen3.5-0.8B MLX in action - demonstration video - coming soon

inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit 是 ModelScope 魔搭社区上的 235.5M 参数text-generation模型，基于 Qwen/Qwen3.5-0.8B 构建。

- **Repository**: inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit
- **Tasks**: text-generation
- **Parameters**: 235.5M
- **Base model**: Qwen/Qwen3.5-0.8B
- **Tags**: quantized, mlx
- **Downloads**: 126
- **Stars**: 0
- **Last updated**: 2026-07-18

Source: https://www.modelscope.cn/models/inferencerlabs/Qwen3.5-0.8B-LM-MLX-9bit

---

**See Qwen3.5-0.8B MLX in action - [demonstration video - coming soon](https://youtu.be/tzF8jv3VGAg)**

#### Tested on a M3 Ultra 512GB RAM using [Inferencer app](https://inferencer.com)
- Single inference ~231.1 tokens/s @ 1000 tokens
- Vision inference: Not included in this language model (LM) only version
- Memory usage: ~0.84 GiB

*q9Bit quant typically achieves near lossless 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%            |
| **q4.8**     | 1.28125    | 93.75%         | 21.15%            |
| **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 - coming soon](https://youtu.be/tzF8jv3VGAg) or visit [Qwen3.5-0.8B](https://huggingface.co/Qwen/Qwen3.5-0.8B).

## 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.
