DeepResearch-9K: A Challenging Benchmark Dataset of Deep-Research Agent

2026-03-01

AI Summary

This work addresses the scarcity of large-scale datasets and open-source training frameworks that reflect real-world complexity for deep research agents. To this end, we introduce DeepResearch-9K, a dataset comprising 9,000 multi-difficulty research-oriented questions, along with DeepResearch-R1, an open-source training framework supporting multi-turn web interaction and reinforcement learning. We propose a cost-effective method to automatically generate high-difficulty search trajectories from multi-hop question answering, augmented with reasoning chains produced by the Tongyi-DeepResearch-30B-A3B model. Leveraging LLM-as-judge and reinforcement learning, our approach trains agents that achieve state-of-the-art performance across multiple deep research benchmarks. Both the dataset and code are publicly released.

Technology Category

• Search and Optimization: Learning to Search
• Multiagent Systems: Adversarial Agents
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Deep-research agents are capable of executing multi-step web exploration, targeted retrieval, and sophisticated question answering. Despite their powerful capabilities, deep-research agents face two critical bottlenecks: (1) the lack of large-scale, challenging datasets with real-world difficulty, and (2) the absence of accessible, open-source frameworks for data synthesis and agent training. To bridge these gaps, we first construct DeepResearch-9K, a large-scale challenging dataset specifically designed for deep-research scenarios built from open-source multi-hop question-answering (QA) datasets via a low-cost autonomous pipeline. Notably, it consists of (1) 9000 questions spanning three difficulty levels from L1 to L3 (2) high-quality search trajectories with reasoning chains from Tongyi-DeepResearch-30B-A3B, a state-of-the-art deep-research agent, and (3) verifiable answers. Furthermore, we develop an open-source training framework DeepResearch-R1 that supports (1) multi-turn web interactions, (2) different reinforcement learning (RL) approaches, and (3) different reward models such as rule-based outcome reward and LLM-as-judge feedback. Finally, empirical results demonstrate that agents trained on DeepResearch-9K under our DeepResearch-R1 achieve state-of-the-art results on challenging deep-research benchmarks. We release the DeepResearch-9K dataset on https://huggingface.co/datasets/artillerywu/DeepResearch-9K and the code of DeepResearch-R1 on https://github.com/Applied-Machine-Learning-Lab/DeepResearch-R1.

Problem

• deep-research agents
• benchmark dataset
• data synthesis
• agent training
• multi-hop QA

Innovation

• Deep-research agent
• benchmark dataset
• multi-hop QA
• reinforcement learning framework
• LLM-as-judge