Rethinking the Value of Generated Tests for LLM Software Engineering Agents

A new study analyzing six large language models on the SWE-bench Verified benchmark found that agent-written tests do not significantly improve issue resolution rates, despite being a common practice. Researchers observed that GPT-5.2 achieved top-tier performance while writing almost no new tests, and that prompt-induced changes in test-writing volume failed to meaningfully alter final outcomes. The findings suggest that current agent-generated testing practices primarily increase process costs and interaction budgets rather than improving task results.

Computer Science Software Engineering Submitted on 8 Feb 2026 v1 https://arxiv.org/abs/2602.07900v1 , last revised 9 Apr 2026 this version, v2 Title:Rethinking the Value of Agent-Generated Tests for LLM-Based Software Engineering Agents View PDF /pdf/2602.07900 HTML experimental https://arxiv.org/html/2602.07900v2 Abstract:Large Language Model LLM code agents increasingly resolve repository-level issues by iteratively editing code, invoking tools, and validating candidate patches. In these workflows, agents often write tests on the fly, but the value of this behavior remains unclear. For example, GPT-5.2 writes almost no new tests yet achieves performance comparable to top-ranking this http URL raises a central question: do such tests meaningfully improve issue resolution, or do they mainly mimic a familiar software-development practice while consuming interaction budget? To better understand the role of agent-written tests, we analyze trajectories produced by six strong LLMs on SWE-bench Verified. Our results show that test writing is common, but resolved and unresolved tasks within the same model exhibit similar test-writing frequencies. When tests are written, they mainly serve as observational feedback channels, with value-revealing print statements appearing much more often than assertion-based checks. Based on these insights, we perform a prompt-intervention study by revising the prompts used with four models to either increase or reduce test writing. The results suggest that prompt-induced changes in the volume of agent-written tests do not significantly change final outcomes in this setting. Taken together, these results suggest that current agent-written testing practices reshape process and cost more than final task outcomes. Submission history From: Zhi Chen view email /show-email/c1d11f2e/2602.07900 Sun, 8 Feb 2026 10:26:31 UTC 372 KB v1 /abs/2602.07900v1 v2 Thu, 9 Apr 2026 13:23:28 UTC 1,239 KB References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .