---
title: Ling-2.6-flash-MLX-Q9
canonical_url: "https://www.modelscope.ai/models/inferencerlabs/Ling-2.6-flash-MLX-Q9"
md_url: "https://www.modelscope.ai/models/inferencerlabs/Ling-2.6-flash-MLX-Q9.md"
repository: inferencerlabs/Ling-2.6-flash-MLX-Q9
last_updated: 2026-06-28
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - bailing_hybrid
architectures:
  - BailingMoeV2_5ForCausalLM
base_model:
  - inclusionAI/Ling-2.6-flash
base_model_relation: quantized
parameters: 32.6B
tensor_type:
  - U32
  - BF16
  - F32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 193
stars: 0
tags:
  - quantized
  - mlx
---

# Ling-2.6-flash-MLX-Q9

> Ling-2.6-flash-MLX-Q9 - An open-source model by inferencerlabs on ModelScope. See Ling-2.6-flash in action - demonstration videos Tested with an M3 Ultra 512 GiB using Inferencer app v1.11.2 Text inference: 48.6 tokens/s @ 1000 tokens 109 GiB (debug build)

inferencerlabs/Ling-2.6-flash-MLX-Q9 is a 32.6B-parameter text-generation model on ModelScope. derived from inclusionAI/Ling-2.6-flash.

- **Repository**: inferencerlabs/Ling-2.6-flash-MLX-Q9
- **Tasks**: text-generation
- **Parameters**: 32.6B
- **Base model**: inclusionAI/Ling-2.6-flash
- **Tags**: quantized, mlx
- **Downloads**: 193
- **Stars**: 0
- **Last updated**: 2026-06-28

Source: https://www.modelscope.ai/models/inferencerlabs/Ling-2.6-flash-MLX-Q9

---

**See Ling-2.6-flash in action - [demonstration videos](https://youtube.com/xcreate)**
   
#### Tested with an M3 Ultra 512 GiB using [Inferencer app v1.11.2](https://inferencer.com)
- Text inference: ~48.6 tokens/s @ 1000 tokens ~109 GiB (debug build)

<p style="margin-bottom:0px;">
<strong>Q9 achieves near lossless accuracy in our coding test</strong>

![Screenshot](https://cdn-uploads.huggingface.co/production/uploads/688479d616f1ec82fa645019/gKvbGEI8hr7raZVqmpR9g.jpeg)
</p>
<table style="border-collapse: collapse; text-align:center; margin-top:10px; margin-bottom:0px;">
<thead>
<tr><th>Quantization (bpw)</th><th>Perplexity</th><th>Token Accuracy</th><th>Missed Divergence</th></tr>
</thead>
<tbody>
<tr><td><strong>Q4.5</strong></td><td>1.0546875</td><td>97.96%</td><td>42.57%</td></tr>
<tr><td><strong>Q6.5</strong></td><td>1.03125</td><td>99.18%</td><td>20.03%</td></tr>
<tr><td><strong>Q9</strong></td><td>1.03125</td><td>99.03%</td><td>18.79%</td></tr>
<tr><td><strong>Base</strong></td><td>1.03125</td><td>100%</td><td>0.000%</td></tr>
</tbody>
</table>

<ul>
<li style="margin:0;">Perplexity: Measures the confidence for predicting base tokens (lower is better)</li>
<li style="margin:0;">Token Accuracy: The percentage of correctly generated base tokens</li>
<li style="margin:0;">Missed Divergence: Measures severity of misses; how much the token was missed by</li>
</ul>

##### Quantized with a modified version of [MLX](https://github.com/ml-explore/mlx)
##### For more details see our [demonstration videos](https://youtube.com/xcreate) or visit [Ling-2.6-flash](https://huggingface.co/inclusionAI/Ling-2.6-flash).

## 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.