Elamur Research⌕
☰
home / papers / 2026 / Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning

Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning

Paper · 2026 · Houxing Ren, Mingjie Zhan, Zimu Lu et al.

Summary

Spreadsheets used in business, auditing and science are often very large. They also carry meaning through layout: multi-level headers, merged cells, several sheets, and colour or shading. Most LLM methods flatten tables into plain text, which loses these cues and can exceed the context window. The authors present SpreadsheetAgent, a two-stage multi-agent system that reads a spreadsheet piece by piece instead of all at once.

In the Structure Extraction Stage, an extraction agent explores the file and builds a structural sketch with row and column summaries in YAML. It uses Python code execution and two helper agents: a vision range agent that renders selected cell ranges as images, and a LaTeX range agent that converts ranges into symbolic tables. Both helpers also check the sketch against the original so that errors do not carry forward. In the Solving Stage, the system answers the task using this sketch and extends it when needed.

On SpreadsheetBench with GPT-OSS-120B, the system beats the ChatGPT Agent baseline. This shows that layout-aware, incremental reading can make spreadsheet QA more robust.

Key points

  • Main result: On SpreadsheetBench with GPT-OSS-120B, SpreadsheetAgent scores 38.16%, compared with 35.27% for ChatGPT Agent (+2.89 points). Evaluation uses SpreadsheetBench and RealHiT Bench.
  • Models: All are open models: Qwen3 (30B, 480B), GPT-OSS (20B, 120B), and GLM-4.5V as the multimodal reasoner.
  • Extractor ablation: A strong extractor helps weaker reasoners most.
    • Qwen3-30B rises from 16.41/12.83 to 22.37/18.42 (soft/hard) with a Qwen3-480B extractor.
    • GPT-OSS-20B rises from 25.26/20.18 to 32.16/29.25 with the same extractor.
    • A weak extractor adds little to a stronger reasoner.
  • Cost: End-to-end extraction takes about 97.49 seconds and 19.54 agent calls per spreadsheet on average, with peak memory of 21 GB.
  • Limitations:
    • The system needs open models that handle both images and function calls, and few such models exist.
    • Distillation into smaller models is left for future work, partly because realistic spreadsheet datasets are scarce.
    • Code: github.com/renhouxing/SpreadsheetAgent.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🏢SensetimeBeijing · China13 works · 2,161 citations🔬Centre for Perceptual and Interactive IntelligenceHong Kong · China8 works · 1,209 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksSoftware · 119 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,685 works

Cite

Cite as Houxing Ren, Mingjie Zhan, Zimu Lu, Ke Wang, Yunqiao Yang, Haotian Hou and Hongsheng Li. 2026. Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.86

@inproceedings{ren2026robust,
  title = {{Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning}},
  author = {Houxing Ren, Mingjie Zhan, Zimu Lu, Ke Wang, Yunqiao Yang, Haotian Hou and Hongsheng Li},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.86},
  url = {https://doi.org/10.18653/v1/2026.acl-long.86}
}

References

35 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lang Cao. 2025. Tablemaster: A recipe to advance table understanding with language models. CoRR, abs/2501.19378. — [arXiv:2501.19378](https://arxiv.org/abs/2501.19378) find
  2. Wenhu Chen, Hongmin Wang, Jianshu Chen, Yunkai Zhang, Hong Wang, Shiyang Li, Xiyou Zhou, and William Yang Wang. 2020. Tabfact: A large-scale dataset for table-based fact verification. In 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. Op…
  3. Yibin Chen, Yifu Yuan, Zeyu Zhang, Yan Zheng, Jinyi Liu, Fei Ni, Jianye Hao, Hangyu Mao, and Fuzheng Zhang. 2025. Sheetagent: Towards a generalist agent for spreadsheet reasoning and manipulation via large language models. In Proceedings of the ACM on Web Conference 2025, WWW 2025, Sydney, NSW, Aust… find
  4. Zhoujun Cheng, Haoyu Dong, Zhiruo Wang, Ran Jia, Jiaqi Guo, Yan Gao, Shi Han, Jian-Guang Lou, and Dongmei Zhang. 2022. Hitab: A hierarchical table dataset for question answering and natural language generation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistic…
  5. DeepSeek-AI, Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, Xiaokang Zhang, Xingkai Yu, Yu Wu, Z. F. Wu, Zhibin Gou, Zhihong Shao, Zhuoshu Li, Ziyi Gao, Aixin Liu, Bing Xue, Bingxuan Wang, Bochao Wu, Bei Feng, Chengda Lu, Chengg… find
  6. Wenyi Hong, Wenmeng Yu, Xiaotao Gu, Guo Wang, Guobing Gan, Haomiao Tang, Jiale Cheng, Ji Qi, Junhui Ji, Lihang Pan, Shuaiqi Duan, Weihan Wang, Yan Wang, Yean Cheng, Zehai He, Zhe Su, Zhen Yang, Ziyang Pan, Aohan Zeng, Baoxu Wang, Boyan Shi, Changyu Pang, Chenhui Zhang, Da Yin, Fan Yang, Guoqing Chen… — GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning
  7. Albert Q. Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de Las Casas, Emma Bou Hanna, Florian Bressand, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Lélio Renard Lavaud, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock,… find
  8. Rihui Jin, Zheyu Xin, Xing Xie, Zuoyi Li, Guilin Qi, Yongrui Chen, Xinbang Dai, Tongtong Wu, and Gholamreza Haffari. 2025. Table-r1: Self-supervised and reinforcement learning for program-based table reasoning in small language models. CoRR, abs/2506.06137. — [arXiv:2506.06137](https://arxiv.org/abs/2506.06137) find
  9. Thomas Joshi, Herman Saini, Neil Dhillon, Antoni Vi-ros i Martin, and Kaoutar El Maghraoui. 2025. Paged attention meets flexattention: Unlocking long-context efficiency in deployed inference. CoRR, abs/2506.07311. find
  10. Yannis Katsis, Saneem A. Chemmengath, Vishwajeet Kumar, Samarth Bharadwaj, Mustafa Canim, Michael R. Glass, Alfio Gliozzo, Feifei Pan, Jaydeep Sen, Karthik Sankaranarayanan, and Soumen Chakrabarti. 2022. AIT-QA: question answering dataset over complex tables in the airline industry. In Proceedings o… find
  11. Hongxin Li, Jingran Su, Yuntao Chen, Qing Li, and Zhaoxiang Zhang. 2023. Sheetcopilot: Bringing software productivity to the next level through large language models. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 202… find
  12. Peng Li, Yeye He, Dror Yashar, Weiwei Cui, Song Ge, Haidong Zhang, Danielle Rifinski Fainman, Dongmei Zhang, and Surajit Chaudhuri. 2024. Table-gpt: Table fine-tuned GPT for diverse table tasks. Proc. ACM Manag. Data, 2(3):176. find
  13. Zeyao Ma, Bohan Zhang, Jing Zhang, Jifan Yu, Xiaokang Zhang, Xiaohan Zhang, Sijia Luo, Xi Wang, and Jie Tang. 2024. Spreadsheetbench: Towards challenging real world spreadsheet manipulation. In Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing S…
  14. OpenAI. 2023. GPT-4 technical report. abs/2303.08774.
  15. CoRR, Panupong Pasupat and Percy Liang. 2015. Compositional semantic parsing on semi-structured tables. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing of the Asian Federation of Natu…
  16. Houxing Ren, Zimu Lu, Weikang Shi, Haotian Hou, Yunqiao Yang, Ke Wang, Aojun Zhou, Junting Pan, Mingjie Zhan, and Hongsheng Li. 2025a. Alignment with fill-in-the-middle for enhancing code generation. find
  17. Linda Wei, Mingjie Zhan, and Hongsheng Li. 2026. From solver to tutor: Evaluating the pedagogical intelligence of llms with kmp-bench. find
  18. Weikang Shi, Aldrich Yu, Rongyao Fang, Houxing Ren, Ke Wang, Aojun Zhou, Changyao Tian, Xinyu Fu, Yuxuan Hu, Zimu Lu, Linjiang Huang, Si Liu, Rui Liu, and Hongsheng Li. 2025. Mathcanvas: Intrinsic visual chain-of-thought for multimodal mathematical reasoning.
  19. Aofeng Su, Aowen Wang, Chao Ye, Chen Zhou, Ga Zhang, Gang Chen, Guangcheng Zhu, Haobo Wang, Haokai Xu, Hao Chen, Haoze Li, Haoxuan Lan, Jiaming Tian, Jing Yuan, Junbo Zhao, Junlin Zhou, Kaizhe Shou, Liangyu Zha, Lin Long, Liyao Li, Pengzuo Wu, Qi Zhang, Qingyi Huang, Saisai Yang, Tao Zhang, Wentao Y… — [arXiv:2411.02059](https://arxiv.org/abs/2411.02059) find
  20. Yuan Sui, Mengyu Zhou, Mingjie Zhou, Shi Han, and Dongmei Zhang. 2024. Table meets LLM: can large language models understand structured table data? A benchmark and empirical study. In Proceedings of the 17th ACM International Conference on Web Search and Data Mining, WSDM 2024, Merida, Mexico, March…
  21. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton-Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, Brian Fuller,… — Llama 2: Open Foundation and Fine-Tuned Chat Models
  22. Houxing Ren, Mingjie Zhan, Zhongyuan Wu, Aojun Zhou, Junting Pan, and Hongsheng Li. 2025b. Re-flectioncoder: Learning from reflection sequence for enhanced one-off code generation. find
  23. Zilong Wang, Hao Zhang, Chun-Liang Li, Julian Martin Eisenschlos, Vincent Perot, Zifeng Wang, Lesly Miculicich, Yasuhisa Fujii, Jingbo Shang, Chen-Yu Lee, and Tomas Pfister. 2024. Chain-of-table: Evolving tables in the reasoning chain for table understanding. In The Twelfth International Conference…
  24. Weikang Shi, Houxing Ren, Junting Pan, Aojun Zhou, Ke Wang, Zimu Lu, Yunqiao Yang, Yuxuan Hu, Pengzuo Wu, Yuhang Yang, Guangcheng Zhu, Chao Ye, Hong Gu, Xu Lu, Ruixuan Xiao, Bowen Bao, Houxing Ren, Mingjie Zhan, Zhongyuan Wu, and Hongsheng Li. 2024. Empowering character-level text infilling by elimi… find
  25. Yijing He, Liangyu Zha, Wentao Ye, Junbo Zhao, and Haobo Wang. 2025a. Realhitbench: A comprehensive realistic hierarchical table benchmark for evaluating llm-based table analysis. In Findings of the Association for Computational Linguistics, ACL 2025, Vienna, Austria, July 27 - August 1, 2025, pages… find
  26. Zhenhe Wu, Jian Yang, Jiaheng Liu, Xianjie Wu, Changzai Pan, Jie Zhang, Yu Zhao, Shuangyong Song, Yongxiang Li, and Zhoujun Li. 2025b. Tabler1: Region-based reinforcement learning for table understanding. CoRR, abs/2505.12415. — [arXiv:2505.12415](https://arxiv.org/abs/2505.12415) find
  27. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jian Yang, Jiaxi Yang, Jingren Zhou, Jun… — Qwen3 Technical Report
  28. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, Keming Lu, Keqin Bao, Kexin Yang, Le Yu, Mei Li, Mingfeng Xue, Pei… find
  29. Yilun Zhao, Lyuhao Chen, Arman Cohan, and Chen Zhao. 2024b. Tapera: Enhancing faithfulness and interpretability in long-form table QA by content planning and execution-based reasoning. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),… find
  30. Mingyu Zheng, Xinwei Feng, Qingyi Si, Qiaoqiao She, Zheng Lin, Wenbin Jiang, and Weiping Wang. 2024. Multimodal table understanding. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok, Thailand, August 11-16, 2024, pa…
  31. Bangbang Zhou, Zuan Gao, Zixiao Wang, Boqiang Zhang, Yuxin Wang, Zhineng Chen, and Hongtao Xie. 2025. Syntab-llava: Enhancing multimodal table understanding with decoupled synthesis. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025, p… find
  32. Ruiyan Zhu, Xi Cheng, Ke Liu, Brian Zhu, Daniel Jin, Neeraj Parihar, Zhoutian Xu, and Oliver Gao. 2025. Sheetmind: An end-to-end llm-powered multi-agent framework for spreadsheet automation. CoRR, abs/2506.12339. — [arXiv:2506.12339](https://arxiv.org/abs/2506.12339) find
  33. Xuanliang Zhang, Dingzirui Wang, Keyan Xu, Qingfu Zhu, and Wanxiang Che. 2025. Rot: Enhancing table reasoning with iterative row-wise traversals. CoRR, abs/2505.15110.
  34. Yunjia Zhang, Jordan Henkel, Avrilia Floratou, Joyce Cahoon, Shaleen Deep, and Jignesh M. Patel. 2024. Reactable: Enhancing react for table question answering. Proc. VLDB Endow., 17(8):1981–1994. find
  35. Weichao Zhao, Hao Feng, Qi Liu, Jingqun Tang, Binghong Wu, Lei Liao, Shu Wei, Yongjie Ye, Hao Liu, Wengang Zhou, Houqiang Li, and Can Huang. 2024a. Tabpedia: Towards comprehensive visual table understanding with concept synergy. In Advances in Neural Information Processing Systems 38: Annual Confere… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ren-2026-robust-real-world-spreadsheet-understanding/