[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85017-en":3,"doc-seo-85017-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85017,1374391975076,"Riley","https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051",8,"Research & Report","Perfopt-Bench: Evaluating Coding Agents on Software Performance Optimization","Coding-agent benchmarks have focused on whether agents generate functionally correct patches, while production software also requires measurable, reproducible speedups on real execution targets. PERFOPT-Bench evaluates software performance optimization as a full agentic loop: profiling executions, diagnosing cross-layer bottlenecks, editing code without correctness regressions, and verifying that gains are not measurement artifacts. Tasks provide correct but suboptimal code and optimize a target metric with hidden tests and trajectory-level audits across long-horizon scenarios.","PERFOPT-Bench: Evaluating Coding Agents on Software Performance  \nOptimization  \nYingyun Cui * OPPO Research Institute [cuiyingyun@oppo.com](cuiyingyun@oppo.com)  \nJiawei Ma†  \nDepartment of Computer Science & Institute of Digital Medicine, City University of Hong Kong [jiaweima@cityu.edu.hk](jiaweima@cityu.edu.hk)  \nYi Xie *  \nUniversity of Arizona [yix@arizona.edu](yix@arizona.edu)  \nBo Liu† University of Arizona [boliu@arizona.edu](boliu@arizona.edu)  \nPiaohong Wang * OPPO Research Institute [wangpiaohong@oppo.com](wangpiaohong@oppo.com)  \nLiangliang Cao†  \nThe Hong Kong Polytechnic University [liangliang.cao@gmail.com](liangliang.cao@gmail.com)  \narXiv :2607 .07744v 1 [ cs . SE] 8 Jul 2026  \nAbstract  \nCoding-agent benchmarks have largely measured whether agents can produce functionally correct patches, but production software also demands measurable speedups on real execution targets. Performance optimization is a distinct agentic task: agents must profile executions, diagnose cross-layer bottlenecks, edit code without breaking correctness, and verify that gains are reproducible rather than measurement artifacts. We introduce PERFOPTBench, a benchmark for evaluating this full performance-engineering loop. Each task provides a correct but deliberately suboptimal codebase and asks the agent to improve a target performance metric; scoring requires hidden correctness tests, verified-speedup measurement, and trajectory-level audit. We evaluate 7 agent stacks with different LLMs and agent frameworks on 12 long-horizon optimization tasks. The results show that optimization performance is workload-dependent rather than determined by model identity alone: no single stack dominates, and changing the agent framework can materially change the same LLM’sper-task speedup profile. We further find that raw speedup is unsafe as a benchmark score, since some large gains arise from benchmarkspecific shortcut exploitation; an exploratory relay pilot suggests that restarting from an externalized optimization summary can recover additional headroom after an initial session stops. The benchmark and our evaluation are available at: [https://anonymous.4open.science/r/](https://anonymous.4open.science/r/)[ ](https://anonymous.4open.science/r/)Dataset-D3CC.  \n1 Introduction  \nCoding agents increasingly operate as autonomous developers: inspecting files, executing tests, and revising code based on feedback (Jimenez et al., 2024 ; Yang et al., 2024 ; Wang et al., 2025 ; Merrill  \n*Equal contribution.  \n†All these authors are corresponding authors.  \net al., 2026) . Benchmarks have followed this shift, moving from function-level synthesis (Chen et al., 2021 ; Austin et al., 2021) to repository repair and terminal environments (Jimenez et al., 2024 ; Jain et al., 2024 ; Zhuo et al., 2025 ; Merrill et al., 2026) . Yet most evaluations still credit agents for passing tests or resolving issues, leaving performance largely outside the scoring target. This omission matters in production settings, where code must remain efficient and stable across hardware, runtime environments, and resource constraints (Williamset al., 2009 ; Nickolls et al., 2008 ; Dean and Barroso, 2013) .  \nSoftware performance tuning differs fundamentally from writing a passing patch. Source-code inspection alone is insufficient; agents must profile real executions, diagnose system-level bottlenecks, and account for platform dependence. A change that accelerates one architecture can slow another, so the central evaluation question shifts from Can the agent produce correct code? to Can it measurably optimize a system on the target hardware? This shift surfaces challenges that single-shot code generation does not. An agent must (i) design and run a profile–patch–verify inner loop without external scaffolding; (ii) hold the optimization objective steady across very long execution trajectories;(iii) organize and compress measurement history when reasoning spans multiple context windows; and","cbCainVpIrbyRwLN","https://ap.wps.com/l/cbCainVpIrbyRwLN","pdf",232689,1,12,"English","en",105,"# Introduction\n## Benchmark design and evaluation focus\n## Contributions and construction pipeline","[{\"question\":\"What is PERFOPT-Bench intended to measure that earlier coding-agent benchmarks often miss?\",\"answer\":\"It measures whether coding agents can measurably optimize software performance on target execution environments, not just produce functionally correct patches.\"},{\"question\":\"How do PERFOPT-Bench tasks define the starting point and success criteria?\",\"answer\":\"Each task starts from a functionally correct but deliberately suboptimal codebase and asks the agent to improve a target performance metric, using hidden correctness tests and verified speedup measurement.\"},{\"question\":\"What do the evaluation results reveal about model identity and agent framework effects?\",\"answer\":\"Optimization performance is workload-dependent; no single agent stack dominates, and changing the agent framework can materially alter the same LLM’s per-task speedup profile.\"}]",1784200312,30,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"perfopt-bench-evaluating-coding-agents-on-software-performance-optimization","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/perfopt-bench-evaluating-coding-agents-on-software-performance-optimization/85017/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is PERFOPT-Bench intended to measure that earlier coding-agent benchmarks often miss?","Question",{"text":75,"@type":76},"It measures whether coding agents can measurably optimize software performance on target execution environments, not just produce functionally correct patches.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do PERFOPT-Bench tasks define the starting point and success criteria?",{"text":80,"@type":76},"Each task starts from a functionally correct but deliberately suboptimal codebase and asks the agent to improve a target performance metric, using hidden correctness tests and verified speedup measurement.",{"name":82,"@type":73,"acceptedAnswer":83},"What do the evaluation results reveal about model identity and agent framework effects?",{"text":84,"@type":76},"Optimization performance is workload-dependent; no single agent stack dominates, and changing the agent framework can materially alter the same LLM’s per-task speedup profile.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":28,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]