Large language models (LLMs) have demonstrated remarkable effectiveness in completing various tasks such as reasoning, translation, and question answering. However the issue of factual incorrect content in LLM-generated responses remains a persistent challenge. In this study, we evaluate both proprietary and open-source LLMs on their ability to respond with relevant research papers and accurate links to articles hosted on the arXiv platform, based on high level prompts. To facilitate this evaluation, we introduce arXivBench, a benchmark specifically designed to assess LLM performance across eight major subject categories on arXiv and five subfields within computer science, one of the most popular categories among them. Our findings reveal a concerning accuracy of LLM-generated responses depending on the subject, with some subjects experiencing significantly lower accuracy than others. Notably, Claude-3.5-Sonnet exhibits a substantial advantage in generating both relevant and accurate responses. And interestingly, most LLMs achieve a much higher accuracy in the Artificial Intelligence sub-field than other sub-fields. This benchmark provides a standardized tool for evaluating the reliability of LLM-generated scientific responses, promoting more dependable use of LLMs in academic and research environments. Our code is open-sourced atthis https URLand our dataset is available on huggingface atthis https URL.
View on arXiv@article{li2025_2504.10496,
title={ ArxivBench: Can LLMs Assist Researchers in Conducting Research? },
author={ Ning Li and Jingran Zhang and Justin Cui },
journal={arXiv preprint arXiv:2504.10496},
year={ 2025 }
}