Haodi Lei (雷皓迪)

figure-cat.png

Nice to meet you! I am a 4th-year undergraduate student at Nanjing University, and also an incoming 2026 Fall Ph.D. student in the School of Artificial Intelligence at Shanghai Jiao Tong University (SJTU) and Shanghai AI Laboratory, advised by Prof. Ning Ding and Prof. Bowen Zhou. I am also fortunate to work closely with Dr. Yafu Li, Dr. Yun Luo, and Dr. Ganqu Cui.

My research interests mainly focus on Reasoning and Efficiency of LLMs, such as Speculative Decoding, Multi-Token Prediction, and how to build strong reasoning systems via reinforcement learning. If you are interested in these topics, feel free to contact me!

News

  • 2026.06   We released Draft-OPD, the first on-policy distillation method for speculative draft models, achieving 5x lossless acceleration for thinking models across diverse tasks.
  • 2026.05   Our team released a 30B-A3B reasoning model that reaches gold-medal level across both physics (IPhO) and math Olympiad (IMO) evaluations, comparable to GPT 5.5, Gemini 3.1 Pro and DeepSeek V3.2 Speciale.
  • 2025.11   Joined Shanghai AI Lab as a research intern.
  • 2025.10   Honored to receive the 2025 National Scholarship of China.
  • 2025.07   I am joining WestLake NLP group as an on-site research intern.

Selected Papers

  1. Preprint
    draft-opd.png
    Draft-OPD: On-Policy Distillation for Speculative Draft Models
    Haodi Lei, Yafu Li, Haoran Zhang, Shunkai Zhang, Qianjia Cheng, Xiaoye Qu, Ganqu Cui, Bowen Zhou, Ning Ding, Yun Luo, and Yu Cheng
    arXiv Code Models Project Page
  2. Preprint
    su01.png
    Achieving Gold-Medal-Level Olympiad Reasoning via Simple and Unified Scaling
    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, and Yu Cheng
    arXiv Code Models Project Page
  3. Preprint
    pi-bench.png
    π-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows
    Haoran Zhang, Luxin Xu, Zhilin Wang, Runquan Gui, Shunkai Zhang, Haodi Lei, Zihao He, Bingsu He, Chicheng Qin, Tong Zhu, Xiaoye Qu, Yang Yang, Yu Cheng, and Yafu Li
    arXiv Code Data Project Page