Hi there! I'm Yuxuan Zhang, a PhD student at the University of British Columbia, advised by Prof. Kelsey R. Allen (Senior Research Scientist, DeepMind). I'm also affiliated with the Vector Institute. I work closely with Prof. Wenhu Chen, Dongfu Jiang, Yubo Wang, and Ping Nie. My research interests span:
I make AI agents work in the real world: ClawBench for live web-agent evaluation (adopted by ByteDance Seed, EMNLP 2026 Findings), RewardHarness for self-evolving agentic post-training (COLM 2026), and VidGround for grounded video understanding.
I'm open to collaborations and research chats. Reach out at
reacher.zh [at] gmail.com ( WeChat works too).
Actively looking for research internships. Let's grab a coffee and have a chat!
ClawBench evaluates AI agents on 153 real-world tasks across 144 live platforms, from booking appointments to filing job applications. It runs on production websites and intercepts only the final submission, keeping evaluation safe without losing real-world complexity.
RewardHarness beats GPT-5 by 5.3 points on image-editing evaluation while using 0.05% of the EditReward data. It reframes reward modeling as context evolution: from 100 preference demonstrations it evolves a library of tools and skills.
COLM 2026
Research Guide Blog Project arXiv Code HuggingFace
@article{zhang2026rewardharness,
title={RewardHarness: Self-Evolving Agentic Post-Training},
author={Zhang, Yuxuan and Du, Penghui and Li, Bo and Wei, Cong and Miao, Junwen and Zhang, Huaisong and Cai, Songcheng and Wang, Yubo and Jiang, Dongfu and Zhang, Yuyu and others},
journal={arXiv preprint arXiv:2605.08703},
year={2026}
}
EMNLP 2026 Findings
Adopted by ByteDance Seed
Research Guide Blog Project arXiv Code Dataset HuggingFace X
@article{zhang2026clawbench,
title={ClawBench: Can AI Agents Complete Everyday Online Tasks?},
author={Zhang, Yuxuan and Wang, Yubo and Zhu, Yipeng and Du, Penghui and Miao, Junwen and Lu, Xuan and Xu, Wendong and Hao, Yunzhuo and Cai, Songcheng and Wang, Xiaochen and others},
journal={arXiv preprint arXiv:2604.08523},
year={2026}
}
Technical Report
Research Guide Project arXiv Code HuggingFace X
@article{zhang2026watch,
title={Watch before you answer: Learning from visually grounded post-training},
author={Zhang, Yuxuan and Hwang, EunJeong and Zhang, Huaisong and Du, Penghui and Jia, Yiming and Jiang, Dongfu and He, Xuan and Zhang, Shenhui and Nie, Ping and West, Peter and others},
journal={arXiv preprint arXiv:2604.05117},
year={2026}
}
Technical Report
@article{wu2026harnessdev,
title={HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?},
author={Wu, Yuhao and Zhang, Jingyuan and Shi, Jiajun and Lei, Xinping and Gu, Qingshui and Zhang, Yuxuan and Wang, Zexuan and He, Chen and Huang, Chen and Song, Maojia and others},
journal={arXiv preprint arXiv:2609.01437},
year={2026}
}
EMNLP 2026 System Demonstrations
Research Guide arXiv Project Code
@article{song2026dr,
title={Dr. Claw: An AI Scientist Workspace for Vibe Research},
author={Song, Dingjie and Zhang, Hanrong and Liu, Dawei and Liu, Yixin and Li, Zongxia and Yuan, Zhengqing and Zhang, Siqi and Zou, Henry Peng and Yan, Zhiling and Zhang, Yuxuan and others},
journal={arXiv preprint arXiv:2609.00365},
year={2026}
}
Technical Report
@article{wu2026aspire,
title={Aspire: Can Models Self-Evolve from Vague Goals?},
author={Wu, Yuhao and Zhang, Jingyuan and Shi, Jiajun and Zhang, Yuxuan and Lei, Xinping and Zhou, Junting and Wang, Zexuan and Wu, Yuchen and Zhou, Huan and Wang, Duo and others},
journal={arXiv preprint arXiv:2608.31111},
year={2026}
}
Technical Report
@article{shi2026s3gym,
title={S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?},
author={Shi, Jiajun and Tao, Siyuan and Wu, Yuhao and Wang, Zexuan and Zhang, Jingyuan and Liu, Jiaheng and Lei, Xinping and Zhang, Xinrong and Fang, Siyuan and Tan, Zhewen and others},
journal={arXiv preprint arXiv:2608.31100},
year={2026}
}
EMNLP 2026
@article{wu2026webworld,
title={WebWorld: The Browser as a World Model for Self-Improving Web Code},
author={Wu, Jiajun and Yang, Jian and Du, Yaxin and Zhang, Wei and Wang, Haowen and Cheng, Junhang and Zhang, Yuxuan and Zheng, Tuney and Liu, Xianglong and Zhou, Ming},
journal={arXiv preprint arXiv:2608.30530},
year={2026}
}
EMNLP 2026
Research Guide arXiv Project HuggingFace
@article{he2026vgi,
title={VGI-BENCH: Probing Visual Intelligence in Video Generation Models},
author={He, Xuan and Wei, Cong and Cheng, Yuhao and Ma, Linrui and Zhang, Yuxuan and Li, Zuojun and Wen, Yuhao and Liu, Zeyi and Hao, Yuren and Cai, Songcheng and others},
journal={arXiv preprint arXiv:2608.19583},
year={2026}
}
Technical Report
@article{fan2026medclaw,
title={MedClaw: Heuristic Agent Harness for Long-Horizon Surgical Video Reasoning},
author={Fan, Yingying and Du, Penghui and Zhu, Leyan and He, Runze and Wu, Zimeng and Zhang, Yuxuan and Chen, Liang and Xie, Jiahao and Wang, Jiangtang and Shao, Shuai and others},
journal={arXiv preprint arXiv:2608.14015},
year={2026}
}
Technical Report
Research Guide arXiv Code Dataset Model HuggingFace
@article{wang2026function,
title={Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models},
author={Wang, Yubo and Liang, Jiarong and Zhang, Yuxuan and Liu, Xuye and Wei, Cong and Zhang, Yuyu and Nie, Ping and Chen, Wenhu},
journal={arXiv preprint arXiv:2607.12463},
year={2026}
}
Technical Report
Research Guide arXiv Code HuggingFace
@article{luo2026learning,
title={Learning from the Self-future: On-policy Self-distillation for dLLMs},
author={Luo, Yifu and Chen, Zeyu and Wang, Haoyu and Hu, Xinhao and Zhang, Yuxuan and Sha, Zhizhou and Liu, Shiwei},
journal={arXiv preprint arXiv:2606.18195},
year={2026}
}
EMNLP 2026
Research Guide arXiv Code HuggingFace
@article{yan2026openskill,
title={OpenSkill: Open-World Self-Evolution for LLM Agents},
author={Yan, Zhiling and Song, Dingjie and Zhang, Hanrong and Liang, Wei and Zhang, Yuxuan and Dai, Yutong and He, Lifang and Yu, Philip S and Xu, Ran and Li, Xiang and others},
journal={arXiv preprint arXiv:2606.06741},
year={2026}
}
Technical Report
Research Guide arXiv HuggingFace
@article{wang2026mira,
title={MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection},
author={Wang, Haowen and Du, Yaxin and Yang, Jian and Wu, Jiajun and Liu, Shukai and Zhang, Yuxuan and Wang, Pingjie and Chen, Siheng and Zheng, Tuney and Zhou, Ming and others},
journal={arXiv preprint arXiv:2605.30288},
year={2026}
}
Technical Report
@article{yao2026dr,
title={Dr. bench: A multidimensional evaluation for deep research agents, from answers to reports},
author={Yao, Yang and Wang, Yixu and Zhang, Yuxuan and Lu, Yi and Gu, Tianle and Li, Lingyu and Zhao, Dingyi and Wu, Keming and Wang, Haozhe and Nie, Ping and others},
year={2026}
}
COLM 2025
Research Guide Project Paper Code Dataset Model Demo
@article{wang2025scholarcopilot,
title={Scholarcopilot: Training large language models for academic writing with accurate citations},
author={Wang, Yubo and Ma, Xueguang and Nie, Ping and Zeng, Huaye and Lyu, Zhiheng and Zhang, Yuxuan and Schneider, Benjamin and Lu, Yi and Yue, Xiang and Chen, Wenhu},
journal={arXiv preprint arXiv:2504.00824},
year={2025}
}
ICLR 2025 Spotlight
Research Guide Project Paper Code
@inproceedings{guan2025retri3d,
title={Retri3D: 3D Neural Graphics Representation Retrieval},
author={Guan, Yushi and Kwan, Daniel and Dandurand, Jean and Yan, Xi and Liang, Ruofan and Zhang, Yuxuan and Jain, Nilesh and Ahuja, Nilesh and Panneer, Selvakumar and Vijaykumar, Nandita},
booktitle={International Conference on Learning Representations},
volume={2025},
pages={31771--31813},
year={2025}
}
TMLR 2026
Research Guide Project Paper Code Dataset Model HF Space Twitter
@article{he2025videoscore2,
title={Videoscore2: Think before you score in generative video evaluation},
author={He, Xuan and Jiang, Dongfu and Nie, Ping and Liu, Minghao and Jiang, Zhengxuan and Su, Mingyi and Ma, Wentao and Lin, Junru and Ye, Chun and Lu, Yi and others},
journal={arXiv preprint arXiv:2509.22799},
year={2025}
}
TMLR 2026
Research Guide Project Paper Code Dataset
@article{yang2025structeval,
title={StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs},
author={Yang, Jialin and Jiang, Dongfu and He, Lipeng and Siu, Sherman and Zhang, Yuxuan and Liao, Disen and Li, Zhuofeng and Zeng, Huaye and Jia, Yiming and Wang, Haozhe and others},
journal={arXiv preprint arXiv:2505.20139},
year={2025}
}