Recent advancements in large language models (LLMs) have significantly advanced the automation of software development tasks, including code synthesis, program repair, and test generation. More recently, researchers and industry practitioners have developed various autonomous LLM agents to perform end-to-end software development tasks. These agents are equipped with the ability to use tools, run commands, observe feedback from the environment, and plan for future actions. However, the complexity of these agent-based approaches, together with the limited abilities of current LLMs, raises the following question: Do we really have to employ complex autonomous software agents? To attempt to answer this question, we build Agentless – an agentless approach to automatically resolve software development issues. Compared to the verbose and complex setup of agent-based approaches, Agentless employs a simplistic three-phase process of localization, repair, and patch validation, without letting the LLM decide future actions or operate with complex tools. Our results on the popular SWE-bench Lite benchmark show that surprisingly the simplistic Agentless is able to achieve both the highest performance (32.67%, 98 correct fixes) and low cost ($0.68) compared with all existing open-source software agents! In fact, Agentless has already been adopted by OpenAI as the go-to approach to showcase the real-world coding performance of both GPT-4o and the new OpenAI o1 models. Furthermore, we manually classified the problems in SWE-bench Lite and found problems with exact ground truth patches or insufficient/misleading issue descriptions. As such, we construct SWE-bench Lite-𝑆 by excluding such problematic issues to perform more rigorous evaluation and comparison. Our work highlights the currently overlooked potential of a simplistic, cost-effective technique in autonomous software development. We hope Agentless will help reset the baseline, starting point, and horizon for autonomous software agents, and inspire future work along this crucial direction.
Wed 25 JunDisplayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change
14:00 - 15:30 | AgentDemonstrations / Industry Papers / Research Papers at Cosmos 3B Chair(s): Chunyang Chen TU Munich | ||
14:00 10mTalk | A Multi-agent Onboarding Assistant based on Large Language Models, Retrieval Augmented Generation, and Chain-of-Thought Demonstrations Andrei Cristian Ionescu Delft University of Technology, Sergey Titov JetBrains Research, Maliheh Izadi Delft University of Technology | ||
14:10 20mTalk | Alibaba LingmaAgent: Improving Automated Issue Resolution via Comprehensive Repository Exploration Industry Papers Yingwei Ma Tongyi Lab, Alibaba, Qingping Yang Tongyi Lab, Alibaba, Rongyu Cao Tongyi Lab, Alibaba, China, Binhua Li Tongyi Lab, Alibaba, China, Fei Huang Tongyi Lab, Alibaba, China, Yongbin Li Tongyi Lab, Alibaba, China | ||
14:30 20mTalk | Demystifying LLM-based Software Engineering Agents Research Papers Chunqiu Steven Xia University of Illinois at Urbana-Champaign, Yinlin Deng University of Illinois at Urbana-Champaign, Soren Dunn University of Illinois Urbana-Champaign, Lingming Zhang University of Illinois at Urbana-Champaign DOI | ||
14:50 20mTalk | AEGIS: An Agent-based Framework for Bug Reproduction from Issue Descriptions Industry Papers Xinchen Wang Harbin Institute of Technology, Pengfei Gao ByteDance, Xiangxin Meng Beihang University, Chao Peng ByteDance, Ruida Hu Harbin Institute of Technology, Shenzhen, Yun Lin Shanghai Jiao Tong University, Cuiyun Gao Harbin Institute of Technology, Shenzhen | ||
15:10 20mTalk | Enhancing Game AI Behaviors with Large Language Models and Agentic AI Industry Papers Ciprian Paduraru Gameloft, and University of Bucharest, Miruna Gabriela Paduraru University of Bucharest
, Alin Stefanescu University of Bucharest |
Cosmos 3B is the second room in the Cosmos 3 wing.
When facing the main Cosmos Hall, access to the Cosmos 3 wing is on the left, close to the stairs. The area is accessed through a large door with the number “3”, which will stay open during the event.