AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents

TL;DR

AgentSwing introduces an adaptive, parallel context management framework for long-horizon web agents, significantly improving efficiency and performance over static methods by dynamically selecting promising context continuations.

Contribution

The paper presents a novel probabilistic framework and a state-aware adaptive routing method that enhances context management in long-horizon web agents.

Findings

AgentSwing outperforms static methods in diverse benchmarks.

Achieves up to 3x fewer interaction turns.

Improves the performance ceiling of long-horizon agents.

Abstract

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.