---
title: MiniMax-M2.1-8bit-gs32
canonical_url: "https://www.modelscope.cn/models/mlx-community/MiniMax-M2.1-8bit-gs32"
md_url: "https://www.modelscope.cn/models/mlx-community/MiniMax-M2.1-8bit-gs32.md"
repository: mlx-community/MiniMax-M2.1-8bit-gs32
last_updated: 2026-02-21
license: other
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - minimax
architectures:
  - MiniMaxM2ForCausalLM
base_model:
  - MiniMaxAI/MiniMax-M2.1
base_model_relation: quantized
parameters: 71.5B
tensor_type:
  - F32
  - BF16
  - U32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 51
stars: 1
tags:
  - transformers
  - mlx
---

# MiniMax-M2.1-8bit-gs32

> MiniMax-M2.1-8bit-gs32 - mlx-community 在 ModelScope 开源的模型。mlx-community/MiniMax-M2.1-8bit-gs32

mlx-community/MiniMax-M2.1-8bit-gs32 是 ModelScope 魔搭社区上的 71.5B 参数text-generation模型，采用 other 许可，基于 MiniMaxAI/MiniMax-M2.1 构建。

- **Repository**: mlx-community/MiniMax-M2.1-8bit-gs32
- **License**: other
- **Tasks**: text-generation
- **Parameters**: 71.5B
- **Base model**: MiniMaxAI/MiniMax-M2.1
- **Tags**: transformers, mlx
- **Downloads**: 51
- **Stars**: 1
- **Last updated**: 2026-02-21

Source: https://www.modelscope.cn/models/mlx-community/MiniMax-M2.1-8bit-gs32

---

# mlx-community/MiniMax-M2.1-8bit-gs32

This model [mlx-community/MiniMax-M2.1-8bit-gs32](https://huggingface.co/mlx-community/MiniMax-M2.1-8bit-gs32) was
converted to MLX format from [MiniMaxAI/MiniMax-M2.1](https://huggingface.co/MiniMaxAI/MiniMax-M2.1)
using mlx-lm version **0.29.1**.


## Recipe:
* 8-bit
* group-size 32
* 9 bits per weight (bpw)

You can find more similar MLX model quants for a single Apple Mac Studio M3 Ultra with 512 GB at https://huggingface.co/bibproj

---

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/MiniMax-M2.1-8bit-gs32")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```