---
title: grug-27b-Text-mlx-8Bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/grug-27b-Text-mlx-8Bit"
md_url: "https://www.modelscope.cn/models/mlx-community/grug-27b-Text-mlx-8Bit.md"
repository: mlx-community/grug-27b-Text-mlx-8Bit
last_updated: 2026-07-23
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3_5
architectures:
  - Qwen3_5ForConditionalGeneration
base_model:
  - ProCreations/grug-27b
base_model_relation: quantized
parameters: 7.6B
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 28
stars: 0
tags:
  - grug
  - reasoning
  - token-efficient
  - agentic
  - tool-use
  - mlx
  - mlx-my-repo
---

# grug-27b-Text-mlx-8Bit

> grug-27b-Text-mlx-8Bit - mlx-community 在 ModelScope 开源的模型。m-i/grug-27b-mlx-8Bit - Text only

mlx-community/grug-27b-Text-mlx-8Bit 是 ModelScope 魔搭社区上的 7.6B 参数text-generation模型，采用 apache-2.0 许可，基于 ProCreations/grug-27b 构建。

- **Repository**: mlx-community/grug-27b-Text-mlx-8Bit
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 7.6B
- **Base model**: ProCreations/grug-27b
- **Tags**: grug, reasoning, token-efficient, agentic, tool-use, mlx, mlx-my-repo
- **Downloads**: 28
- **Stars**: 0
- **Last updated**: 2026-07-23

Source: https://www.modelscope.cn/models/mlx-community/grug-27b-Text-mlx-8Bit

---

# m-i/grug-27b-mlx-8Bit _-_ Text only

The Model [m-i/grug-27b-mlx-8Bit](https://huggingface.co/m-i/grug-27b-mlx-8Bit) was converted to MLX format from [ProCreations/grug-27b](https://huggingface.co/ProCreations/grug-27b) using mlx-lm version **0.31.2**, therefore, no vision.

## Use with mlx

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

```python
from mlx_lm import load, generate

model, tokenizer = load("m-i/grug-27b-mlx-8Bit")

prompt="hello"

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

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