AgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research Agents

TL;DR

AgentDisCo introduces a disentangled, collaborative architecture for deep research agents, improving exploration, exploitation, and report synthesis through iterative refinement and meta-optimization.

Contribution

It presents a novel agentic framework with disentangled modules and meta-optimization, enabling self-refinement and personalized research recommendations.

Findings

Achieved performance comparable to or surpassing leading closed-source systems on established benchmarks.

Introduced GALA, a new benchmark mining user interests from browsing behavior.

Developed a rendering agent for converting reports into visual posters.

Abstract

In this paper, we present AgentDisCo, a novel Disentangled and Collaborative agentic architecture that formulates deep research as an adversarial optimization problem between information exploration and exploitation. Unlike existing approaches that conflate these two processes into a single module, AgentDisCo employs a critic agent to evaluate generated outlines and refine search queries, and a generator agent to retrieve updated results and revise outlines accordingly. The iteratively refined outline is then passed to a downstream report writer that synthesizes a comprehensive research report. The overall workflow supports both handcrafted and automatically discovered design strategies via a meta-optimization harness, in which the generator agent is repurposed as a scoring agent to evaluate critic outputs and generate quality signals. Powerful code-generation agents (e.g., Claude-Code, Codex) systematically explore agent configurations and construct a policy bank, a structured repository of reusable design strategies, enabling the framework to self-refine without extensive human intervention. We evaluate AgentDisCo on three established deep research benchmarks (DeepResearchBench, DeepConsult, DeepResearchGym) using Gemini-2.5-Pro, achieving performance comparable to or surpassing leading closed-source systems. Observing that existing benchmarks inadequately reflect real-world user needs, we introduce GALA (General AI Life Assistants), a benchmark that mines latent research interests from users' historical browsing behavior. We further develop a rendering agent that converts research reports into visually rich poster presentations, and demonstrate an end-to-end product, AutoResearch Your Interest, which delivers personalized deep research recommendations derived from individual browsing histories.