通过简单统一的扩展实现奥赛金牌级推理
本文提出一种将预训练推理模型转化为严格奥赛求解器的统一方法。该方法首先采用反向困惑度课程进行监督微调，以灌输严谨的证明搜索与自我检查行为；随后通过两阶段强化学习流程扩展这些能力，最终结合测试时扩展提升性能。基于此方案训练的30B参数模型SU-01，在仅使用约34万条短轨迹微调和200步强化学习后，能稳定处理超过10万token的长轨迹难题，并在IMO、USAMO、IPhO等数学与物理奥赛中达到金牌级表现，同时展现出向数学物理之外科学领域的强推理泛化能力。
IMO 金牌级推理模型又多了一个，SU-01 的方法干净统一，特别在超长推理链上的稳定性是真正突破，做推理模型训练和竞赛级 AI 的可以认真读一下。
Authors:Yafu Li, Runzhe Zhan, Haoran Zhang, Shunkai Zhang, Yizhuo Li, Zhilin Wang, Jiacheng Chen, Futing Wang, Xuyang Hu, Yuchen Fan, Bangjie Xu, Yucheng Su, Xinmiao Han, Chenxi Li, Haodi Lei, Yufeng Zhao, Zejin Lin, Qianjia Cheng, Tong Zhu, Xiaoye Qu, Ganqu Cui, Peng Ye, Yun Luo, Zhouchen Lin, Yu Qiao, Bowen Zhou, Ning Ding, Yu Cheng
Abstract:Recent progress in reasoning models has substantially advanced long-horizon mathematical and scientific problem solving, with several systems now reaching gold-medal-level performance on International Mathematical Olympiad (IMO) and International Physics Olympiad (IPhO) problems. In this paper, we introduce a simple and unified recipe for converting a post-trained reasoning backbone into a rigorous olympiad-level solver. The recipe first uses a reverse-perplexity curriculum for SFT to instill rigorous proof-search and self-checking behaviors, then scales these behaviors through a two-stage RL pipeline that progresses from RL with verifiable rewards to more delicate proof-level RL, and finally boosts solving performance with test-time scaling. Applying this recipe, we train a 30B-A3B backbone with SFT on around 340K sub-8K-token trajectories followed by 200 RL steps. The resulting model, SU-01, supports stable reasoning on difficult problems with trajectories exceeding 100K tokens, while achieving gold-medal-level performance on mathematical and physical olympiad competitions, including IMO 2025/USAMO 2026 and IPhO 2024/2025. It also demonstrates strong generalization of scientific reasoning to domains beyond mathematics and physics.
| Comments: | Technical Report. 77 pages |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2605.13301 [cs.AI] |
| (or arXiv:2605.13301v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.13301 arXiv-issued DOI via DataCite |
Submission history
From: Runzhe Zhan [view email]
[v1]
Wed, 13 May 2026 10:13:26 UTC (638 KB)
来源：HuggingFace Daily Papers（社区热门论文） · arxiv.org