---
license: apache-2.0
base_model:
- Qwen/Qwen2.5-Math-1.5B
library_name: transformers
pipeline_tag: text-generation
tags:
- mathematical-reasoning
- reinforcement-learning
- grpo
---

# Model Card for PrAg-PO Qwen2.5-Math-1.5B step2480

**If you are using the model, a star to our [github repo](https://github.com/wenquanlu/PrAg-PO) would be really appreciated! 😊**

**This checkpoint achieves 80+ on MATH500!**

This is the step 2480 checkpoint when training Qwen2.5-Math-1.5B on MATH Level-3-to-5 Dataset using PrAg-PO. The training procedure is outlined in the paper [PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning](https://arxiv.org/abs/2602.03190).

### Model Sources

- **Repository 🤖:** [https://github.com/wenquanlu/PrAg-PO](https://github.com/wenquanlu/PrAg-PO)
- **Paper 📝:** [PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning](https://arxiv.org/abs/2602.03190)

## Uses

This model is intended for mathematical reasoning tasks. It leverages prompt augmentation to generate reasoning traces under diverse templates, increasing rollout diversity and stability during RL training.

### Results

<p align="center">
<img src="https://raw.githubusercontent.com/wenquanlu/prompt-augmentation-GRPO/master/imgs/table_result.png">
</p>


## Citation
```bibtex
@misc{lu2026pragpopromptaugmentedpolicy,
      title={PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning}, 
      author={Wenquan Lu and Hai Huang and Enqi Liu and Randall Balestriero},
      journal={arXiv preprint arXiv:2602.03190},
      url={https://arxiv.org/abs/2602.03190},
      year={2026},
}
```