DeepResearchBench
Emerging4papers using it
2026first seen
The 'DeepResearchBench' is a dataset/benchmark used to evaluate the performance of deep research agents in generating and refining research outlines and reports through adversarial optimization between information exploration and exploitation.
Papers using DeepResearchBench (4)
- Co-ReAct: Rubrics as Step-Level Collaborators for ReAct AgentsAgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research AgentsSearch More, Think Less: Rethinking Long-Horizon Agentic Search for Efficiency and GeneralizationSelf-Manager: Parallel Agent Loop for Long-form Deep Research