Wuying-Browser-Agent: Real-World Centric Fundamental Long-Horizon Browser Agents
- URL: http://arxiv.org/abs/2608.17319v1
- Date: Tue, 18 Aug 2026 03:23:31 GMT
- Title: Wuying-Browser-Agent: Real-World Centric Fundamental Long-Horizon Browser Agents
- Abstract summary: Wuying-Browser-Agent is a unified framework that addresses execution, supervision, optimization, and evaluation.<n>A structured browser harness provides stable execution primitives and decision-oriented context management.<n> BrowserBench is a benchmark of 350 tasks averaging 37.9 steps.
- Score: 34.59625581903402
- License: http://creativecommons.org/licenses/by-nc-sa/4.0/
- Abstract: Browser agents perform well on short, clean demonstrations, but real deployment is fundamentally different: agents must sustain dozens of decisions on live websites while recovering from mistakes and navigating complex UIs. We argue that closing this gap requires alignment at every level of the pipeline, including execution, supervision, optimization, and evaluation, rather than scale alone. We present Wuying-Browser-Agent, a unified framework that addresses each of these levels. A structured browser harness provides stable execution primitives and decision-oriented context management. Reflection and UI-specialized Curriculum SFT (RUIC-SFT) explicitly trains on recovery trajectories and complex-UI interactions. Divergence-Aware Online GRPO (DAO-GRPO) improves long-horizon credit assignment through potential-based reward shaping and divergence-aware step weighting. Finally, we introduce BrowserBench, a bilingual real-web benchmark of 350 tasks averaging 37.9 steps, because most existing benchmarks are too short to expose long-horizon failure modes. Wuying-Browser-Agent-27B achieves 80.6\% on WebVoyager, 66.7\% on Online-Mind2Web, and 65.1\% on BrowserBench, establishing a new open-source state of the art on browser-use benchmarks. The same pipeline also transfers beyond browser use, demonstrating strong general agentic ability and reaching an average score of 73.8 on Tau2-Bench, Claw-Eval, and BFCL-v4.
Related papers
- BrowserForge: Scaling Web Episode via Parallel Browser Sandboxes [54.96895469970198]
BrowserForge generates web interaction data at scale by driving many browser sandboxes in parallel over the open web.<n> BrowserForge couples three components: an open-web sourcing stage that exposes the agent to hundreds of thousands of real, openly reachable websites.<n>A rule-plus-model cleaning pipeline removes failed runs and rewrites the surviving reasoning into a single unified chain-of-thought style.
arXiv Detail & Related papers (2026-08-25T17:35:42Z) - CAP: A Scalable Benchmark for Evaluating Cross-Site Browser Agents with Complex Actions and Perception [22.620664364045286]
Large language models are increasingly deployed as autonomous agents that interact with the web through browsers.<n>We introduce CAP, a scalable benchmark for evaluating browser agents on cross-site, human-like web tasks.<n>We construct 420 tasks across 108 real-world websites and 24 domains under careful quality control.
arXiv Detail & Related papers (2026-08-09T01:08:19Z) - Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments [67.2234316079859]
GauntletBench is a web-based benchmark for evaluating agent generalisation in challenging scenarios.<n>It focuses on three underexplored capabilities (temporal perception, graphical understanding, and 3D reasoning) across five less-covered professional applications.<n>Our empirical results reveal that frontier agentic systems remain far from achieving human-level performance.
arXiv Detail & Related papers (2026-06-12T12:32:24Z) - GTA: Generating Long-Horizon Tasks for Web Agents at Scale [82.43869456830664]
We introduce a scalable framework, GTA, that integrates crawling, retrieval-based seeding, in-context generation, and automated quality control.<n>We instantiate the pipeline on over 50 websites covering e-commerce, government, forums, and news, with multilingual and multi-hop coverage.<n>Our contributions are three-fold: (i) formalizing multi-hop web-agent task generation, (ii) proposing an efficient and validated pipeline for automatic data creation, and (iii) releasing a dynamic benchmark with reproducible evaluation.
arXiv Detail & Related papers (2026-05-28T01:05:50Z) - WebForge: Breaking the Realism-Reproducibility-Scalability Trilemma in Browser Agent Benchmark [10.793973999774026]
Existing browser agent benchmarks face a fundamental trilemma: real-website benchmarks lack due to content drift, controlled environments sacrifice realism by omitting real-web noise, and both require costly manual curation that limits scalability.<n>We present WebForge, the first fully automated framework that resolves this trilemma through a four-agent pipeline -- Plan, Generate, Refine, and.<n>A seven-dimensional difficulty control framework structures task design along navigation depth, visual complexity, reasoning difficulty, and more, enabling systematic capability profiling beyond single aggregate scores.
arXiv Detail & Related papers (2026-04-13T04:45:27Z) - TimeWarp: Evaluating Web Agents by Revisiting the Past [7.017865728670461]
We introduce TimeWarp, a benchmark that emulates the evolving web using containerized environments that vary in UI, design, and layout.<n>Our experiments reveal web agents' vulnerability to changes and the limitations of behavior cloning (BC) on single-version trajectories.<n>We propose TimeTraj, a simple yet effective algorithm that uses plan distillation to collect trajectories across multiple versions.
arXiv Detail & Related papers (2026-03-05T08:43:06Z) - WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions [14.49822563708463]
Training web agents to navigate complex, real-world websites requires them to master $itsubtasks$.<n>We introduce WARC-Bench, a novel web navigation benchmark featuring 438 tasks designed to evaluate multimodal AI agents on subtasks.<n>We show that WARC-Bench is challenging for leading computer-use models, with the highest observed success rate being 64.8%.
arXiv Detail & Related papers (2025-10-10T21:20:51Z) - BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks [51.803138848305814]
We introduce BrowserArena, a live open-web agent evaluation platform that collects user-submitted tasks.<n>We identify three consistent failure modes: captcha resolution, pop-up banner removal, and direct navigation to URLs.<n>Our findings surface both the diversity and brittleness of current web agents.
arXiv Detail & Related papers (2025-10-02T15:22:21Z) - The BrowserGym Ecosystem for Web Agent Research [151.90034093362343]
BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents.<n>We propose an extended BrowserGym-based ecosystem for web agent research, which unifies existing benchmarks from the literature.<n>We conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across 6 popular web agent benchmarks.
arXiv Detail & Related papers (2024-12-06T23:43:59Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.