SciResearcher: Scaling Deep Research Agents for Frontier Scientific Reasoning

TL;DR

SciResearcher is an automated framework that constructs scientific reasoning data, enabling the development of advanced AI agents for frontier scientific discovery and achieving state-of-the-art performance on multiple benchmarks.

Contribution

It introduces a novel automated data construction paradigm for scientific reasoning, leading to the development of SciResearcher-8B, a new state-of-the-art scientific agent model.

Findings

SciResearcher-8B achieves 19.46% on HLE-Bio/Chem-Gold benchmark.

It surpasses several larger proprietary agents in scientific reasoning tasks.

The framework improves performance on SuperGPQA-Hard-Biology and TRQA-Literature benchmarks.

Abstract

Frontier scientific reasoning is rapidly emerging as a key foundation for advancing AI agents in automated scientific discovery. Deep research agents offer a promising approach to this challenge. These models develop robust problem-solving capabilities through post-training on information-seeking tasks, which are typically curated via knowledge graph construction or iterative web browsing. However, these strategies face inherent limitations in frontier science, where domain-specific knowledge is scattered across sparse and heterogeneous academic sources, and problem solving requires sophisticated computation and reasoning far beyond factual recall. To bridge this gap, we introduce SciResearcher, a fully automated agentic framework for frontier-science data construction. SciResearcher synthesizes diverse conceptual and computational tasks grounded in academic evidence, while eliciting information acquisition, tool-integrated reasoning, and long-horizon capabilities. Leveraging the curated data for supervised fine-tuning and agentic reinforcement learning, we develop SciResearcher-8B, an agent foundation model that achieves 19.46% on the HLE-Bio/Chem-Gold benchmark, establishing a new state of the art at its parameter scale and surpassing several larger proprietary agents. It further achieves 13-15% absolute gains on SuperGPQA-Hard-Biology and TRQA-Literature benchmarks. Overall, SciResearcher introduces a new paradigm for automated data construction for frontier scientific reasoning and offers a scalable path toward future scientific agents.