PATCHEVAL: A New Benchmark for Evaluating LLMs on Patching Real-World Vulnerabilities

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wei, Zichao, Zeng, Jun, Wen, Ming, Yu, Zeliang, Cheng, Kai, Zhu, Yiding, Guo, Jingyi, Zhou, Shiqi, Yin, Le, Su, Xiaodong, Ma, Zhechao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911265783283712
author Wei, Zichao
Zeng, Jun
Wen, Ming
Yu, Zeliang
Cheng, Kai
Zhu, Yiding
Guo, Jingyi
Zhou, Shiqi
Yin, Le
Su, Xiaodong
Ma, Zhechao
author_facet Wei, Zichao
Zeng, Jun
Wen, Ming
Yu, Zeliang
Cheng, Kai
Zhu, Yiding
Guo, Jingyi
Zhou, Shiqi
Yin, Le
Su, Xiaodong
Ma, Zhechao
contents Software vulnerabilities are increasing at an alarming rate. However, manual patching is both time-consuming and resource-intensive, while existing automated vulnerability repair (AVR) techniques remain limited in effectiveness. Recent advances in large language models (LLMs) have opened a new paradigm for AVR, demonstrating remarkable progress. To examine the capability of LLMs in AVR, several vulnerability benchmarks have been proposed recently. However, they still suffer from key limitations of outdated vulnerabilities, limited language coverage, unreliable patch validation, and insufficient reproducibility. To overcome these challenges, we introduce PATCHEVAL, a multilingual benchmark for Go, JavaScript, and Python, languages for which existing benchmarks remain unexplored. PATCHEVAL curates a dataset of 1,000 vulnerabilities drawn from CVEs reported between 2015 and 2025, covering 65 distinct CWEs. A subset of 230 CVEs is further equipped with runtime sandbox environments, enabling patch verification through both security tests and functionality tests. To provide a systematic comparison of LLM-based vulnerability repair, we evaluate a series of state-of-the-art LLMs and agents, presenting an in-depth analysis that empirically yields key insights to guide future research in AVR.
format Preprint
id arxiv_https___arxiv_org_abs_2511_11019
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PATCHEVAL: A New Benchmark for Evaluating LLMs on Patching Real-World Vulnerabilities
Wei, Zichao
Zeng, Jun
Wen, Ming
Yu, Zeliang
Cheng, Kai
Zhu, Yiding
Guo, Jingyi
Zhou, Shiqi
Yin, Le
Su, Xiaodong
Ma, Zhechao
Cryptography and Security
Software Engineering
Software vulnerabilities are increasing at an alarming rate. However, manual patching is both time-consuming and resource-intensive, while existing automated vulnerability repair (AVR) techniques remain limited in effectiveness. Recent advances in large language models (LLMs) have opened a new paradigm for AVR, demonstrating remarkable progress. To examine the capability of LLMs in AVR, several vulnerability benchmarks have been proposed recently. However, they still suffer from key limitations of outdated vulnerabilities, limited language coverage, unreliable patch validation, and insufficient reproducibility. To overcome these challenges, we introduce PATCHEVAL, a multilingual benchmark for Go, JavaScript, and Python, languages for which existing benchmarks remain unexplored. PATCHEVAL curates a dataset of 1,000 vulnerabilities drawn from CVEs reported between 2015 and 2025, covering 65 distinct CWEs. A subset of 230 CVEs is further equipped with runtime sandbox environments, enabling patch verification through both security tests and functionality tests. To provide a systematic comparison of LLM-based vulnerability repair, we evaluate a series of state-of-the-art LLMs and agents, presenting an in-depth analysis that empirically yields key insights to guide future research in AVR.
title PATCHEVAL: A New Benchmark for Evaluating LLMs on Patching Real-World Vulnerabilities
topic Cryptography and Security
Software Engineering
url https://arxiv.org/abs/2511.11019