{"ID":2856248,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.11221","arxiv_id":"2510.11221","title":"WebRouter: Query-specific Router via Variational Information Bottleneck for Cost-sensitive Web Agent","abstract":"LLM-brained web agents offer powerful capabilities for web automation but face a critical cost-performance trade-off. The challenge is amplified by web agents' inherently complex prompts that include goals, action histories, and environmental states, leading to degraded LLM ensemble performance. To address this, we introduce WebRouter, a novel query-specific router trained from an information-theoretic perspective. Our core contribution is a cost-aware Variational Information Bottleneck (ca-VIB) objective, which learns a compressed representation of the input prompt while explicitly penalizing the expected operational cost. Experiments on five real-world websites from the WebVoyager benchmark show that WebRouter reduces operational costs by a striking 87.8\\% compared to a GPT-4o baseline, while incurring only a 3.8\\% accuracy drop.","short_abstract":"LLM-brained web agents offer powerful capabilities for web automation but face a critical cost-performance trade-off. The challenge is amplified by web agents' inherently complex prompts that include goals, action histories, and environmental states, leading to degraded LLM ensemble performance. To address this, we int...","url_abs":"https://arxiv.org/abs/2510.11221","url_pdf":"https://arxiv.org/pdf/2510.11221v1","authors":"[\"Tao Li\",\"Jinlong Hu\",\"Yang Wang\",\"Junfeng Liu\",\"Xuejun Liu\"]","published":"2025-10-13T10:05:43Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"Large Language Model\"]","has_code":false}