SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agents
Published in Technical Report, 2025
Recommended citation: Xuan-Phi Nguyen, Shrey Pandit, Revanth Gangi Reddy, Austin Xu, Silvio Savarese, Caiming Xiong, Shafiq Joty (2025). SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agents. Technical Report.
Paper Link: https://arxiv.org/abs/2509.06283
Abstract
This work develops a single reasoning agent trained with reinforcement learning to conduct multi-step research using external tools.