---
title: longcat-audiodit-3.5b-8bit-mlx
canonical_url: "https://www.modelscope.cn/models/mlx-community/longcat-audiodit-3.5b-8bit-mlx"
md_url: "https://www.modelscope.cn/models/mlx-community/longcat-audiodit-3.5b-8bit-mlx.md"
repository: mlx-community/longcat-audiodit-3.5b-8bit-mlx
last_updated: 2026-04-16
license: mit
pipeline_tag: text-to-speech
tasks:
  - text-to-speech
model_type:
  - audiodit
base_model:
  - meituan-longcat/LongCat-AudioDiT-3.5B
base_model_relation: quantized
parameters: 1.2B
tensor_type:
  - U32
  - F32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - zh
  - en
downloads: 24
stars: 0
tags:
  - mlx
  - tts
  - speech
  - longcat
  - audiodit
  - diffusion
  - quantized
  - int8
  - apple-silicon
---

# longcat-audiodit-3.5b-8bit-mlx

> longcat-audiodit-3.5b-8bit-mlx - mlx-community 在 ModelScope 开源的模型。LongCat AudioDiT 3.5B — MLX 8-bit

mlx-community/longcat-audiodit-3.5b-8bit-mlx 是 ModelScope 魔搭社区上的 1.2B 参数text-to-speech模型，采用 mit 许可，基于 meituan-longcat/LongCat-AudioDiT-3.5B 构建。

- **Repository**: mlx-community/longcat-audiodit-3.5b-8bit-mlx
- **License**: mit
- **Tasks**: text-to-speech
- **Parameters**: 1.2B
- **Base model**: meituan-longcat/LongCat-AudioDiT-3.5B
- **Tags**: mlx, tts, speech, longcat, audiodit, diffusion, quantized, int8, apple-silicon
- **Downloads**: 24
- **Stars**: 0
- **Last updated**: 2026-04-16

Source: https://www.modelscope.cn/models/mlx-community/longcat-audiodit-3.5b-8bit-mlx

---

# LongCat AudioDiT 3.5B — MLX 8-bit

This repository contains a self-contained MLX-native int8 conversion of
LongCat AudioDiT 3.5B for local text-to-speech on Apple Silicon.

> Note
> This repo is a community mirror of the canonical MLX conversion maintained by
> [AppAutomaton](https://github.com/appautomaton) at
> [`appautomaton/longcat-audiodit-3.5b-8bit-mlx`](https://huggingface.co/appautomaton/longcat-audiodit-3.5b-8bit-mlx).

## Model Details

- Developed by: AppAutomaton
- Shared by: `mlx-community`
- Original MLX repo: [`appautomaton/longcat-audiodit-3.5b-8bit-mlx`](https://huggingface.co/appautomaton/longcat-audiodit-3.5b-8bit-mlx)
- Upstream model: [`meituan-longcat/LongCat-AudioDiT-3.5B`](https://huggingface.co/meituan-longcat/LongCat-AudioDiT-3.5B)
- Task: text-to-speech
- Runtime: MLX on Apple Silicon
- Precision: int8 quantized weights with bundled tokenizer

## Bundle Contents

This bundle is self-contained and includes:

- `config.json`
- `model.safetensors`
- tokenizer files (`tokenizer.json`, `tokenizer_config.json`, `special_tokens_map.json`)

## How to Get Started

Command-line generation with [`mlx-speech`](https://github.com/appautomaton/mlx-speech):

```bash
python scripts/generate/longcat_audiodit.py \
  --text "Hello from LongCat AudioDiT." \
  --model-dir /path/to/longcat-audiodit-3.5b-8bit-mlx \
  --output-audio outputs/longcat.wav
```

Voice cloning:

```bash
python scripts/generate/longcat_audiodit.py \
  --text "Hello from LongCat AudioDiT." \
  --prompt-text "Original speaker text." \
  --prompt-audio /path/to/prompt.wav \
  --model-dir /path/to/longcat-audiodit-3.5b-8bit-mlx \
  --output-audio outputs/longcat_clone.wav \
  --guidance-method apg
```

Minimal Python usage:

```python
from pathlib import Path

from mlx_speech.generation.longcat_audiodit import generate_longcat_audiodit

result = generate_longcat_audiodit(
    text="Hello from LongCat AudioDiT.",
    model_dir=Path("/path/to/longcat-audiodit-3.5b-8bit-mlx"),
    output_audio="outputs/longcat.wav",
)
```

## Notes

- This repo contains the quantized MLX runtime artifact only.
- The conversion preserves the LongCat AudioDiT diffusion transformer and
  bundled VAE for waveform decode.
- The current bundle is intended for local MLX runtime use and parity validation.
- This mirror is a duplicated repo, not an automatically synchronized namespace mirror.

## Links

- Canonical MLX repo: [`appautomaton/longcat-audiodit-3.5b-8bit-mlx`](https://huggingface.co/appautomaton/longcat-audiodit-3.5b-8bit-mlx)
- Source code: [`mlx-speech`](https://github.com/appautomaton/mlx-speech)
- More examples: [AppAutomaton](https://github.com/appautomaton)

## License

MIT License — following the upstream license published with
[`meituan-longcat/LongCat-AudioDiT-3.5B`](https://huggingface.co/meituan-longcat/LongCat-AudioDiT-3.5B).