xbench-DeepSearch

The 'xbench-DeepSearch' dataset/benchmark contains synthesized samples used to evaluate the performance of search agents, particularly in the context of complex, multi-hop reasoning tasks.

Papers using xbench-DeepSearch (4)

OpenSeeker: Democratizing Frontier Search Agents by Fully Open-Sourcing Training Data2026MiroFlow: Towards High-Performance and Robust Open-Source Agent Framework for General Deep Research Tasks2026WebLeaper: Empowering Efficiency and Efficacy in WebAgent via Enabling Info-Rich Seeking2025Multi-Agent Deep Research: Training Multi-Agent Systems with M-GRPO2025