---
title: Qwen3.8-27B-MTP-4bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/Qwen3.8-27B-MTP-4bit"
md_url: "https://www.modelscope.cn/models/mlx-community/Qwen3.8-27B-MTP-4bit.md"
repository: mlx-community/Qwen3.8-27B-MTP-4bit
last_updated: 2026-11-11
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3_5_mtp
base_model:
  - Qwen/Qwen3.8-27B
base_model_relation: quantized
parameters: 66.4M
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 1777
stars: 1
tags:
  - mlx
  - mlx-vlm
  - qwen3_5_mtp
  - qwen3.8
  - qwen3.8-27b
  - qwen
  - mtp
  - speculative-decoding
  - draft-model
  - 4-bit
---

# Qwen3.8-27B-MTP-4bit

> Qwen3.8-27B-MTP-4bit - mlx-community 在 ModelScope 开源的模型。Qwen3.8-27B-MTP-4bit

mlx-community/Qwen3.8-27B-MTP-4bit 是 ModelScope 魔搭社区上的 66.4M 参数text-generation模型，采用 apache-2.0 许可，基于 Qwen/Qwen3.8-27B 构建。

- **Repository**: mlx-community/Qwen3.8-27B-MTP-4bit
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 66.4M
- **Base model**: Qwen/Qwen3.8-27B
- **Tags**: mlx, mlx-vlm, qwen3_5_mtp, qwen3.8, qwen3.8-27b, qwen, mtp, speculative-decoding, draft-model, 4-bit
- **Downloads**: 1777
- **Stars**: 1
- **Last updated**: 2026-11-11

Source: https://www.modelscope.cn/models/mlx-community/Qwen3.8-27B-MTP-4bit

---

# Qwen3.8-27B-MTP-4bit

This repository contains the Multi-Token Prediction (MTP) drafter weights split from [`Qwen/Qwen3.8-27B`](https://huggingface.co/Qwen/Qwen3.8-27B) and quantized with `mlx_vlm.convert` for `mlx-vlm` speculative decoding.

This is not a standalone model. Use it as the draft model alongside a compatible Qwen3.8 27B target checkpoint.

## Use with mlx-vlm

```bash
mlx_vlm generate \
  --model mlx-community/Qwen3.8-27B-4bit \
  --draft-model mlx-community/Qwen3.8-27B-MTP-4bit \
  --prompt "Write a quicksort in Python." \
  --max-tokens 256 \
  --enable-thinking
```

`--draft-kind mtp` is detected automatically from `model_type`.

## Model details

- Model type: `qwen3_5_mtp`
- MTP block size: 3
- Target architecture: Qwen3.8 27B
- Precision: MLX affine 4-bit, group size 64
- Runtime: MLX / `mlx-vlm`
- Source revision: `1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0`

The adapter holds only the MTP drafter weights; the target supplies the token embeddings and language-model head at runtime. Use an adapter and target derived from the same Qwen3.8 27B checkpoint. The upstream Apache 2.0 license and model limitations continue to apply.