AACR-Bench: Evaluating Automatic Code Review with Holistic Repository-Level Context

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Lei, Yu, Yongda, Yu, Minghui, Guo, Xinxin, Zhuang, Zhengqi, Rong, Guoping, Shao, Dong, Shen, Haifeng, Kuang, Hongyu, Li, Zhengfeng, Wang, Boge, Zhang, Guoan, Xiang, Bangyu, Xu, Xiaobin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917234332401664
author Zhang, Lei
Yu, Yongda
Yu, Minghui
Guo, Xinxin
Zhuang, Zhengqi
Rong, Guoping
Shao, Dong
Shen, Haifeng
Kuang, Hongyu
Li, Zhengfeng
Wang, Boge
Zhang, Guoan
Xiang, Bangyu
Xu, Xiaobin
author_facet Zhang, Lei
Yu, Yongda
Yu, Minghui
Guo, Xinxin
Zhuang, Zhengqi
Rong, Guoping
Shao, Dong
Shen, Haifeng
Kuang, Hongyu
Li, Zhengfeng
Wang, Boge
Zhang, Guoan
Xiang, Bangyu
Xu, Xiaobin
contents High-quality evaluation benchmarks are pivotal for deploying Large Language Models (LLMs) in Automated Code Review (ACR). However, existing benchmarks suffer from two critical limitations: first, the lack of multi-language support in repository-level contexts, which restricts the generalizability of evaluation results; second, the reliance on noisy, incomplete ground truth derived from raw Pull Request (PR) comments, which constrains the scope of issue detection. To address these challenges, we introduce AACR-Bench a comprehensive benchmark that provides full cross-file context across multiple programming languages. Unlike traditional datasets, AACR-Bench employs an "AI-assisted, Expert-verified" annotation pipeline to uncover latent defects often overlooked in original PRs, resulting in a 285% increase in defect coverage. Extensive evaluations of mainstream LLMs on AACR-Bench reveal that previous assessments may have either misjudged or only partially captured model capabilities due to data limitations. Our work establishes a more rigorous standard for ACR evaluation and offers new insights on LLM based ACR, i.e., the granularity/level of context and the choice of retrieval methods significantly impact ACR performance, and this influence varies depending on the LLM, programming language, and the LLM usage paradigm e.g., whether an Agent architecture is employed. The code, data, and other artifacts of our evaluation set are available at https://github.com/alibaba/aacr-bench .
format Preprint
id arxiv_https___arxiv_org_abs_2601_19494
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AACR-Bench: Evaluating Automatic Code Review with Holistic Repository-Level Context
Zhang, Lei
Yu, Yongda
Yu, Minghui
Guo, Xinxin
Zhuang, Zhengqi
Rong, Guoping
Shao, Dong
Shen, Haifeng
Kuang, Hongyu
Li, Zhengfeng
Wang, Boge
Zhang, Guoan
Xiang, Bangyu
Xu, Xiaobin
Software Engineering
Artificial Intelligence
High-quality evaluation benchmarks are pivotal for deploying Large Language Models (LLMs) in Automated Code Review (ACR). However, existing benchmarks suffer from two critical limitations: first, the lack of multi-language support in repository-level contexts, which restricts the generalizability of evaluation results; second, the reliance on noisy, incomplete ground truth derived from raw Pull Request (PR) comments, which constrains the scope of issue detection. To address these challenges, we introduce AACR-Bench a comprehensive benchmark that provides full cross-file context across multiple programming languages. Unlike traditional datasets, AACR-Bench employs an "AI-assisted, Expert-verified" annotation pipeline to uncover latent defects often overlooked in original PRs, resulting in a 285% increase in defect coverage. Extensive evaluations of mainstream LLMs on AACR-Bench reveal that previous assessments may have either misjudged or only partially captured model capabilities due to data limitations. Our work establishes a more rigorous standard for ACR evaluation and offers new insights on LLM based ACR, i.e., the granularity/level of context and the choice of retrieval methods significantly impact ACR performance, and this influence varies depending on the LLM, programming language, and the LLM usage paradigm e.g., whether an Agent architecture is employed. The code, data, and other artifacts of our evaluation set are available at https://github.com/alibaba/aacr-bench .
title AACR-Bench: Evaluating Automatic Code Review with Holistic Repository-Level Context
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2601.19494