AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915897367592960 |
|---|---|
| author | Feng, Zhaopeng Su, Liangcai Zhang, Zhen Wang, Xinyu Zhang, Xiaotian Wang, Xiaobin Fang, Runnan Zhang, Qi Li, Baixuan Cai, Shihao Ye, Rui Chen, Hui Yong, Jiang Zhou, Joey Tianyi Qian, Chenxiong Xie, Pengjun Hooi, Bryan Liu, Zuozhu Zhou, Jingren |
| author_facet | Feng, Zhaopeng Su, Liangcai Zhang, Zhen Wang, Xinyu Zhang, Xiaotian Wang, Xiaobin Fang, Runnan Zhang, Qi Li, Baixuan Cai, Shihao Ye, Rui Chen, Hui Yong, Jiang Zhou, Joey Tianyi Qian, Chenxiong Xie, Pengjun Hooi, Bryan Liu, Zuozhu Zhou, Jingren |
| contents | As large language models (LLMs) evolve into autonomous agents for long-horizon information-seeking, managing finite context capacity has become a critical bottleneck. Existing context management methods typically commit to a single fixed strategy throughout the entire trajectory. Such static designs may work well in some states, but they cannot adapt as the usefulness and reliability of the accumulated context evolve during long-horizon search. To formalize this challenge, we introduce a probabilistic framework that characterizes long-horizon success through two complementary dimensions: search efficiency and terminal precision. Building on this perspective, we propose AgentSwing, a state-aware adaptive parallel context management routing framework. At each trigger point, AgentSwing expands multiple context-managed branches in parallel and uses lookahead routing to select the most promising continuation. Experiments across diverse benchmarks and agent backbones show that AgentSwing consistently outperforms strong static context management methods, often matching or exceeding their performance with up to $3\times$ fewer interaction turns while also improving the ultimate performance ceiling of long-horizon web agents. Beyond the empirical gains, the proposed probabilistic framework provides a principled lens for analyzing and designing future context management strategies for long-horizon agents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_27490 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents Feng, Zhaopeng Su, Liangcai Zhang, Zhen Wang, Xinyu Zhang, Xiaotian Wang, Xiaobin Fang, Runnan Zhang, Qi Li, Baixuan Cai, Shihao Ye, Rui Chen, Hui Yong, Jiang Zhou, Joey Tianyi Qian, Chenxiong Xie, Pengjun Hooi, Bryan Liu, Zuozhu Zhou, Jingren Computation and Language Artificial Intelligence Multiagent Systems As large language models (LLMs) evolve into autonomous agents for long-horizon information-seeking, managing finite context capacity has become a critical bottleneck. Existing context management methods typically commit to a single fixed strategy throughout the entire trajectory. Such static designs may work well in some states, but they cannot adapt as the usefulness and reliability of the accumulated context evolve during long-horizon search. To formalize this challenge, we introduce a probabilistic framework that characterizes long-horizon success through two complementary dimensions: search efficiency and terminal precision. Building on this perspective, we propose AgentSwing, a state-aware adaptive parallel context management routing framework. At each trigger point, AgentSwing expands multiple context-managed branches in parallel and uses lookahead routing to select the most promising continuation. Experiments across diverse benchmarks and agent backbones show that AgentSwing consistently outperforms strong static context management methods, often matching or exceeding their performance with up to $3\times$ fewer interaction turns while also improving the ultimate performance ceiling of long-horizon web agents. Beyond the empirical gains, the proposed probabilistic framework provides a principled lens for analyzing and designing future context management strategies for long-horizon agents. |
| title | AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents |
| topic | Computation and Language Artificial Intelligence Multiagent Systems |
| url | https://arxiv.org/abs/2603.27490 |