paper-with-me

Papers

LoCoBench-Agent: An Interactive Benchmark for LLM Agents in Long-Context Software Engineering

2025-11-17 · Jielin Qiu, Zuxin Liu, Zhiwei Liu, Rithesh Murthy, Jianguo Zhang, Haolin Chen, Shiyu Wang, Ming Zhu, Liangwei Yang, Juntao Tan, Roshan Ram, Akshara Prabhakar, Tulika Awalgaonkar, Zixiang Chen, Zhepeng Cen, Cheng Qian, Shelby Heinecke, Weiran Yao, Silvio Savarese, Caiming Xiong, Huan Wang arxiv

As large language models (LLMs) evolve into sophisticated autonomous agents capable of complex software development tasks, evaluating their real-world capabilities becomes critical. While existing benchmarks like LoCoBench~\cite{qiu2025locobench} assess long-context code understanding, they focus on single-turn evaluation and cannot capture the multi-turn interactive nature, tool usage patterns, and adaptive reasoning required by real-world coding agents. We introduce \textbf{LoCoBench-Agent}, a comprehensive evaluation framework specifically designed to assess LLM agents in realistic, long-context software engineering workflows. Our framework extends LoCoBench's 8,000 scenarios into interactive agent environments, enabling systematic evaluation of multi-turn conversations, tool usage efficiency, error recovery, and architectural consistency across extended development sessions. We also introduce an evaluation methodology with 9 metrics across comprehension and efficiency dimensions. Our framework provides agents with 8 specialized tools (file operations, search, code analysis) and evaluates them across context lengths ranging from 10K to 1M tokens, enabling precise assessment of long-context performance. Through systematic evaluation of state-of-the-art models, we reveal several key findings: (1) agents exhibit remarkable long-context robustness; (2) comprehension-efficiency trade-off exists with negative correlation, where thorough exploration increases comprehension but reduces efficiency; and (3) conversation efficiency varies dramatically across models, with strategic tool usage patterns differentiating high-performing agents. As the first long-context LLM agent benchmark for software engineering, LoCoBench-Agent establishes a rigorous foundation for measuring agent capabilities, identifying performance gaps, and advancing autonomous software development at scale.

📄 PDF Abstract BibTeX arXiv:2511.13998

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

LoCoBench: A Benchmark for Long-Context Large Language Models in Complex Software Engineering

2025-09-11 · Jielin Qiu, Zuxin Liu, Zhiwei Liu, Rithesh Murthy 외 arxiv

The emergence of long-context language models with context windows extending to millions of tokens has created new opportunities for sophisticated code understanding and software development evaluation. We propose LoCoBe…

Long-Context Understanding

One Interaction Is Worth a Thousand Guesses: Benchmarking the Interactive Capabilities of Deep Research Agents

2026-01-10 · Yingchaojie Feng, Qiang Huang, Xiaoya Xie, Zhaorui Yang 외 arxiv

Deep research agents powered by Large Language Models (LLMs) can perform multi-step reasoning, web exploration, and long-form report generation. However, existing systems remain largely autonomous, assuming fully specifi…

LifelongAgentBench: Evaluating LLM Agents as Lifelong Learners

2025-05-17 · Junhao Zheng, Xidi Cai, Qiuke Li, Duzhen Zhang 외

Lifelong learning is essential for intelligent agents operating in dynamic environments. Current large language model (LLM)-based agents, however, remain stateless and unable to accumulate or transfer knowledge over time…

Language ModelingLanguage ModellingLarge Language ModelLifelong learning

Commit0: Library Generation from Scratch

2024-12-02 · Wenting Zhao, Nan Jiang, Celine Lee, Justin T Chiu 외

With the goal of benchmarking generative systems beyond expert software development ability, we introduce Commit0, a benchmark that challenges AI agents to write libraries from scratch. Agents are provided with a specifi…

BenchmarkingCode Generation

TextQuests: How Good are LLMs at Text-Based Video Games?

2025-07-31 · Long Phan, Mantas Mazeika, Andy Zou, Dan Hendrycks arxiv

Evaluating AI agents within complex, interactive environments that mirror real-world challenges is critical for understanding their practical capabilities. While existing agent benchmarks effectively assess skills like t…