Research Paper (preprint): DeepScholar-Bench: A Live Benchmark and Automated Evaluation for Generative Research Synthesis

The preprint linked below was recently posted on arXiv.

Title

DeepScholar-Bench: A Live Benchmark and Automated Evaluation for Generative Research Synthesis

Authors

Liana Patel
Stanford University

Negar Arabzadeh
UC Berkeley

Harshit Gupta
Stanford University

Ankita Sundar
UC Berkeley

Ion Stoica
UC Berkeley

Matei Zaharia
UC Berkeley

Carlos Guestrin
Stanford University

Source

via arXiv
DOI: 10.48550/arXiv.2508.20033

Abstract

The ability to research and synthesize knowledge is central to human expertise and progress. An emerging class of systems promises these exciting capabilities through generative research synthesis, performing retrieval over the live web and synthesizing discovered sources into long-form, cited summaries. However, evaluating such systems remains an open challenge: existing question-answering benchmarks focus on short-form factual responses, while expert-curated datasets risk staleness and data contamination. Both fail to capture the complexity and evolving nature of real research synthesis tasks. In this work, we introduce DeepScholar-bench, a live benchmark and holistic, automated evaluation framework designed to evaluate generative research synthesis. DeepScholar-bench draws queries from recent, high-quality ArXiv papers and focuses on a real research synthesis task: generating the related work sections of a paper by retrieving, synthesizing, and citing prior research. Our evaluation framework holistically assesses performance across three key dimensions, knowledge synthesis, retrieval quality, and verifiability. We also develop DeepScholar-base, a reference pipeline implemented efficiently using the LOTUS API. Using the DeepScholar-bench framework, we perform a systematic evaluation of prior open-source systems, search AI’s, OpenAI’s DeepResearch, and DeepScholar-base. We find that DeepScholar-base establishes a strong baseline, attaining competitive or higher performance than each other method. We also find that DeepScholar-bench remains far from saturated, with no system exceeding a score of  across all metrics. These results underscore the difficulty of DeepScholar-bench, as well as its importance for progress towards AI systems capable of generative research synthesis. We make our code available at https://github.com/guestrin-lab/deepscholar-bench.

Direct to Abstract + Link to Full Text

Filed under: Data Files, Journal Articles, News