RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users

Suyu Ye, Haojun Shi, Darren Shih, Hyokun Yun, Tanya G. Roosta, Tianmin Shu · 2025

Read RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users by Suyu Ye, Haojun Shi, Darren Shih, Hyokun Yun, Tanya G. Roosta, Tianmin Shu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-10445.txt.

Description

Suyu Ye and co-authors built RealWebAssist in 2025, a benchmark for AI help with long web tasks from real users. Real help means following a chain of instructions over time, not one tidy task. Human instructions can be vague, can call for varied levels of help, and can shift as the user's mind changes.

The benchmark gathers chains of instructions from real-world users, each telling a web agent what to do next. It tests visual grounding on screen layouts and the reading of ambiguous requests in context. That design sets it apart from older web agent tests built on single, clear tasks. Readers get a harder, more honest test of web agents. A system that passes it is closer to helping with actual errands, not just demos.

Catalog record

CREATORSuyu Ye, Haojun Shi, Darren Shih, Hyokun Yun, Tanya G. Roosta, Tianmin Shu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-10445

Browse the full library · About the Tower of Babel