The Year in AI Papers
The year/Independent research

Paper 2509.13309

WebResearcher: Unleashing unbounded reasoning capability in Long-Horizon Agents

Zile Qiao, Guoxin Chen, Xuanzhong Chen, Donglei Yu, Wenbiao Yin, Xinyu Wang, Zhen Zhang, Baixuan Li, Huifeng Yin, Kuan Li, Rui Min, Minpeng Liao, et al.

Published
Sep 2025
Research lab
Independent
Citations
66
GitHub
Not linked

01 In brief

Summary

WebResearcher is a framework for deep-research agents that addresses the limitations of mono-contextual approaches, which suffer from context suffocation and noise contamination.

It introduces IterResearch, an iterative paradigm that reformulates deep research as a Markov Decision Process, where agents periodically consolidate findings into evolving reports and reconstruct focused workspaces.

WebFrontier, a scalable data synthesis engine, generates high-quality training data through tool-augmented complexity escalation.

At inference, a Research-Synthesis Framework enables parallel multi-agent exploration and integrative synthesis.

Evaluated on 6 benchmarks, WebResearcher achieves state-of-the-art performance, surpassing proprietary systems: 36.7% on HLE, 51.7% on BrowseComp, 75.7% on GAIA, and 85.1% on Frames.

Ablations confirm the iterative paradigm's importance, and training data from this paradigm also improves mono-contextual methods.

The framework scales with parallel research, with n=8 offering a good trade-off between performance and cost.

02 From the paper

Abstract

Recent advances in deep-research systems have demonstrated the potential for AI agents to autonomously discover and synthesize knowledge from external sources. In this paper, we introduce WebResearcher, a novel framework for building such agents through two key components: (1) WebResearcher, an iterative deep-research paradigm that reformulates deep research as a Markov Decision Process, where agents periodically consolidate findings into evolving reports while maintaining focused workspaces, overcoming the context suffocation and noise contamination that plague existing mono-contextual approaches; and (2) WebFrontier, a scalable data synthesis engine that generates high-quality training data through tool-augmented complexity escalation, enabling systematic creation of research tasks that bridge the gap between passive knowledge recall and active knowledge construction. Notably, we find that the training data from our paradigm significantly enhances tool-use capabilities even for traditional mono-contextual methods. Furthermore, our paradigm naturally scales through parallel thinking, enabling concurrent multi-agent exploration for more comprehensive conclusions. Extensive experiments across 6 challenging benchmarks demonstrate that WebResearcher achieves state-of-the-art performance, even surpassing frontier proprietary systems.