@inproceedings{zhang-etal-2026-datascibench,
title = "{D}ata{S}ci{B}ench: An {LLM} Agent Benchmark for Data Science",
author = "Zhang, Dan and
Zhoubian, Sining and
Cai, Min and
Li, Fengzu and
Yang, Lekang and
Wang, Wei and
Dong, Tianjiao and
Hu, Ziniu and
Tang, Jie and
Yue, Yisong",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.181/",
pages = "3685--3728",
ISBN = "979-8-89176-395-1",
abstract = "This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Models (LLMs) in data science. Unlike existing benchmarks limited to single task, simple evaluation metrics, and readily available ground truth (GT), DataSciBench is built on curated, natural, and challenging prompts with complex evaluation criteria and uncertain GT. To bridge the gap, we develop a semi-automated GT generation pipeline, integrating LLM-based self-consistency and human verification to ensure accuracy, predefined task types, and aggregate functions (metrics). Furthermore, we introduce an innovative Intention-Function-Code (IFC) framework, assessing code execution outcomes through metrics and programmatic rules. Evaluating 26 models (8 API-based, 8 open-source general, 9 code generation, and 1 agentic models), our approach offers rigorous insights into LLM strengths and weaknesses. Experimental results show API-based models outperform open-source counterparts across all metrics, with DeepAnalyze-8B leading among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-datascibench">
<titleInfo>
<title>DataSciBench: An LLM Agent Benchmark for Data Science</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sining</namePart>
<namePart type="family">Zhoubian</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Min</namePart>
<namePart type="family">Cai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fengzu</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lekang</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianjiao</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ziniu</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yisong</namePart>
<namePart type="family">Yue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Models (LLMs) in data science. Unlike existing benchmarks limited to single task, simple evaluation metrics, and readily available ground truth (GT), DataSciBench is built on curated, natural, and challenging prompts with complex evaluation criteria and uncertain GT. To bridge the gap, we develop a semi-automated GT generation pipeline, integrating LLM-based self-consistency and human verification to ensure accuracy, predefined task types, and aggregate functions (metrics). Furthermore, we introduce an innovative Intention-Function-Code (IFC) framework, assessing code execution outcomes through metrics and programmatic rules. Evaluating 26 models (8 API-based, 8 open-source general, 9 code generation, and 1 agentic models), our approach offers rigorous insights into LLM strengths and weaknesses. Experimental results show API-based models outperform open-source counterparts across all metrics, with DeepAnalyze-8B leading among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench.</abstract>
<identifier type="citekey">zhang-etal-2026-datascibench</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.181/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>3685</start>
<end>3728</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DataSciBench: An LLM Agent Benchmark for Data Science
%A Zhang, Dan
%A Zhoubian, Sining
%A Cai, Min
%A Li, Fengzu
%A Yang, Lekang
%A Wang, Wei
%A Dong, Tianjiao
%A Hu, Ziniu
%A Tang, Jie
%A Yue, Yisong
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F zhang-etal-2026-datascibench
%X This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Models (LLMs) in data science. Unlike existing benchmarks limited to single task, simple evaluation metrics, and readily available ground truth (GT), DataSciBench is built on curated, natural, and challenging prompts with complex evaluation criteria and uncertain GT. To bridge the gap, we develop a semi-automated GT generation pipeline, integrating LLM-based self-consistency and human verification to ensure accuracy, predefined task types, and aggregate functions (metrics). Furthermore, we introduce an innovative Intention-Function-Code (IFC) framework, assessing code execution outcomes through metrics and programmatic rules. Evaluating 26 models (8 API-based, 8 open-source general, 9 code generation, and 1 agentic models), our approach offers rigorous insights into LLM strengths and weaknesses. Experimental results show API-based models outperform open-source counterparts across all metrics, with DeepAnalyze-8B leading among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench.
%U https://aclanthology.org/2026.findings-acl.181/
%P 3685-3728
Markdown (Informal)
[DataSciBench: An LLM Agent Benchmark for Data Science](https://aclanthology.org/2026.findings-acl.181/) (Zhang et al., Findings 2026)
ACL
- Dan Zhang, Sining Zhoubian, Min Cai, Fengzu Li, Lekang Yang, Wei Wang, Tianjiao Dong, Ziniu Hu, Jie Tang, and Yisong Yue. 2026. DataSciBench: An LLM Agent Benchmark for Data Science. In Findings of the Association for Computational Linguistics: ACL 2026, pages 3685–3728, San Diego, California, United States. Association for Computational Linguistics.