We present PCL-Reasoner-V1.5, a 32-billion-parameter large language model (LLM) for mathematical reasoning. The model is built upon Qwen2.5-32B and refined via supervised fine-tuning (SFT) followed by reinforcement learning (RL). A central innovation is our proposed offline RL method, which provides superior training stability and efficiency over standard online RL methods such as GRPO. Our model achieves state-of-the-art performance among models post-trained on Qwen2.5-32B, attaining average accuracies of 90.9% on AIME 2024 and 85.6% on AIME 2025. Our work demonstrates offline RL as a stable and efficient paradigm for advancing reasoning in LLMs. All experiments were conducted on Huawei Ascend 910C NPUs.
@article{arxiv.2601.14716,
title = {PCL-Reasoner-V1.5: Advancing Math Reasoning with Offline Reinforcement Learning},
author = {Yao Lu and Dengdong Fan and Jianzheng Nie and Fan Xu and Jie Chen and Bin Zhou and Yonghong Tian},
journal= {arXiv preprint arXiv:2601.14716},
year = {2026}
}