---
title: MathGLM
canonical_url: "https://www.modelscope.cn/models/ZhipuAI/MathGLM"
md_url: "https://www.modelscope.cn/models/ZhipuAI/MathGLM.md"
repository: ZhipuAI/MathGLM
last_updated: 2026-01-20
license: afl-3.0
pipeline_tag: text-generation
tasks:
  - text-generation
library_name:
  - pytorch
frameworks:
  - pytorch
downloads: 7359
stars: 28
---

# MathGLM

> MathGLM - ZhipuAI 在 ModelScope 开源的模型。MathGLM 在数据集上对 GLM-10B 进行了微调，并添加了额外的多步算术运算和文本中描述的数学问题，在 5,000 个样本的中国数学问题测试集上实现了与 GPT-4 类似的性能。

ZhipuAI/MathGLM 是 ModelScope 魔搭社区上的text-generation模型，采用 afl-3.0 许可。

- **Repository**: ZhipuAI/MathGLM
- **License**: afl-3.0
- **Tasks**: text-generation
- **Downloads**: 7359
- **Stars**: 28
- **Last updated**: 2026-01-20

Source: https://www.modelscope.cn/models/ZhipuAI/MathGLM

---

## Model description
MathGLM-10B is finetuned from GLM-10B on a dataset with additional multi-step arithmetic operations and math problems described in text, achieves similar performance to GPT-4 on a 5,000-samples Chinese math problem
test set.


## How to use 
First, you shoud run the following command to pip sat.
```
pip install SwissArmyTransformer
```
Second, you run the inference code to evaluate our MathGLM-10B.
```
bash inference.sh
```


## Citation
Please cite our paper if you find this code useful for your research:
```
@article{yang2023gpt,
  title={GPT Can Solve Mathematical Problems Without a Calculator},
  author={Yang, Zhen and Ding, Ming and Lv, Qingsong and Jiang, Zhihuan and He, Zehai and Guo, Yuyi and Bai, Jinfeng and Tang, Jie},
  journal={arXiv preprint arXiv:2309.03241},
  year={2023}
}
```