---
title: Trinity-Large-Thinking-5bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/Trinity-Large-Thinking-5bit"
md_url: "https://www.modelscope.cn/models/mlx-community/Trinity-Large-Thinking-5bit.md"
repository: mlx-community/Trinity-Large-Thinking-5bit
last_updated: 2026-04-03
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - afmoe
architectures:
  - AfmoeForCausalLM
base_model:
  - arcee-ai/Trinity-Large-Thinking
base_model_relation: quantized
parameters: 74.7B
tensor_type:
  - BF16
  - U32
  - F32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
  - es
  - fr
  - de
  - it
  - pt
  - ru
  - ar
  - hi
  - ko
  - zh
downloads: 8
stars: 0
tags:
  - reasoning
  - agentic
  - tool-calling
  - thinking
  - mlx
---

# Trinity-Large-Thinking-5bit

> Trinity-Large-Thinking-5bit - mlx-community 在 ModelScope 开源的模型。mlx-community/Trinity-Large-Thinking-5bit

mlx-community/Trinity-Large-Thinking-5bit 是 ModelScope 魔搭社区上的 74.7B 参数text-generation模型，采用 apache-2.0 许可，基于 arcee-ai/Trinity-Large-Thinking 构建。

- **Repository**: mlx-community/Trinity-Large-Thinking-5bit
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 74.7B
- **Base model**: arcee-ai/Trinity-Large-Thinking
- **Tags**: reasoning, agentic, tool-calling, thinking, mlx
- **Downloads**: 8
- **Stars**: 0
- **Last updated**: 2026-04-03

Source: https://www.modelscope.cn/models/mlx-community/Trinity-Large-Thinking-5bit

---

# mlx-community/Trinity-Large-Thinking-5bit

This model [mlx-community/Trinity-Large-Thinking-5bit](https://huggingface.co/mlx-community/Trinity-Large-Thinking-5bit) was
converted to MLX format from [arcee-ai/Trinity-Large-Thinking](https://huggingface.co/arcee-ai/Trinity-Large-Thinking)
using mlx-lm version **0.31.2**.

## Use with mlx

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

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Trinity-Large-Thinking-5bit")

prompt = "hello"

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

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