---
title: Ling-3.1-flash-1019-oQ2.5e-mtp
canonical_url: "https://www.modelscope.cn/models/mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp"
md_url: "https://www.modelscope.cn/models/mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp.md"
repository: mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp
last_updated: 2026-10-28
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - bailing_hybrid
architectures:
  - BailingMoeV3ForCausalLM
base_model:
  - inclusionAI/Ling-3.1-flash-1019
base_model_relation: quantized
parameters: 65.0B
tensor_type:
  - F32
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 9
stars: 0
tags:
  - mlx
  - oq
  - quantized
  - bailing_hybrid
  - mtp
---

# Ling-3.1-flash-1019-oQ2.5e-mtp

> Ling-3.1-flash-1019-oQ2.5e-mtp - mlx-community 在 ModelScope 开源的模型。Ling-3.1-flash-1019-oQ2.5e-mtp

mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp 是 ModelScope 魔搭社区上的 65.0B 参数text-generation模型，采用 mit 许可，基于 inclusionAI/Ling-3.1-flash-1019 构建。

- **Repository**: mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 65.0B
- **Base model**: inclusionAI/Ling-3.1-flash-1019
- **Tags**: mlx, oq, quantized, bailing_hybrid, mtp
- **Downloads**: 9
- **Stars**: 0
- **Last updated**: 2026-10-28

Source: https://www.modelscope.cn/models/mlx-community/Ling-3.1-flash-1019-oQ2.5e-mtp

---

# Ling-3.1-flash-1019-oQ2.5e-mtp

This model was quantized using [oQ](https://github.com/jundot/omlx) (oMLX v0.7.3rc1) mixed-precision quantization.

A mixed-precision MLX checkpoint of [inclusionAI/Ling-3.1-flash-1019](https://huggingface.co/inclusionAI/Ling-3.1-flash-1019), an upgraded version of [Ling-3.1-flash](https://huggingface.co/inclusionAI/Ling-3.1-flash) with improved reasoning and agentic capabilities. It is a hybrid-linear mixture-of-experts model with 560B total and 25B active parameters, built from 63 KDA linear-attention layers and 9 gated MLA layers, with 512 routed experts of which 8 are active per token. It includes the language model and the MTP (NextN) layer, which oMLX runs as Lightning MTP speculative decoding.

The effective average is 3.132 bits/weight for the complete checkpoint.

## Quantization and Bit Distribution

oQ2.5e allocates bits by measured layer sensitivity, then rounds each group with an importance matrix (256 samples x 512 tokens, collected layer by layer from the original FP8 weights) and a weighted least-squares refit of its scale and bias.

| Storage format | Logical weights | Share of total | Tensor storage | Effective bits/weight |
|---|---|---|---|---|
| Affine 2-bit, group 64 | 300.648B | 52.550% | 87.500 GiB | 2.50 |
| Affine 3-bit, group 64 | 246.961B | 43.166% | 100.625 GiB | 3.50 |
| Affine 4-bit, group 64 | 8.053B | 1.408% | 4.219 GiB | 4.50 |
| Affine 8-bit, group 128 | 1.085B | 0.190% | 1.042 GiB | 8.25 |
| Affine 8-bit, group 64 | 15.333B | 2.680% | 15.173 GiB | 8.50 |
| BF16 | 0.041B | 0.007% | 0.076 GiB | 16.00 |
| FP32 | 0.000B | 0.000% | 0.000 GiB | 32.00 |
| **Total** | **572.121B** | **100%** | **208.635 GiB** | **3.132** |

## Storage Breakdown

| Component | Logical weights | Storage |
|---|---|---|
| Language backbone | 563.946B | 219.331 GB / 204.267 GiB |
| MTP head | 8.174B | 4.690 GB / 4.367 GiB |
| **Total** | **572.121B** | **224.020 GB / 208.635 GiB** |

## License

MIT, following the source model.