{"id":"https://openalex.org/W7163668543","doi":"https://doi.org/10.48550/arxiv.2606.05670","title":"Do More Agents Help? Controlled and Protocol-Aligned Evaluation of LLM Agent Workflows","display_name":"Do More Agents Help? Controlled and Protocol-Aligned Evaluation of LLM Agent Workflows","publication_year":2026,"publication_date":"2026-06-04","ids":{"openalex":"https://openalex.org/W7163668543","doi":"https://doi.org/10.48550/arxiv.2606.05670"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.05670","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.05670","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.05670","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137968969","display_name":"Yuhang Fu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fu, Yuhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109637392","display_name":"Ruishan Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Ruishan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138000997","display_name":"Jiaqi Shao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Jiaqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137947699","display_name":"Huiyu Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Huiyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137995757","display_name":"Zhengtao Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Zhengtao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137981005","display_name":"Bing Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Bing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137970204","display_name":"Tao Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Tao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11986","display_name":"Scientific Computing and Data Management","score":0.22519999742507935,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T11986","display_name":"Scientific Computing and Data Management","score":0.22519999742507935,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10215","display_name":"Semantic Web and Ontologies","score":0.1298000067472458,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10456","display_name":"Multi-Agent Systems and Negotiation","score":0.09870000183582306,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/workflow","display_name":"Workflow","score":0.9193000197410583},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.705299973487854},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.4643000066280365},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.38339999318122864},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.3199000060558319}],"concepts":[{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.9193000197410583},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.705299973487854},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6909000277519226},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.4643000066280365},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.38339999318122864},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.3783999979496002},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.3199000060558319},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.303600013256073},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.29649999737739563},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.26739999651908875},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.2572999894618988},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.25290000438690186},{"id":"https://openalex.org/C41550386","wikidata":"https://www.wikidata.org/wiki/Q529909","display_name":"Multi-agent system","level":2,"score":0.25220000743865967}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.05670","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.05670","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.05670","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.05670","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Does":[0],"adding":[1],"more":[2,111,132],"agents":[3],"help":[4],"an":[5,28],"LLM":[6],"workflow":[7,123],"once":[8],"compared":[9],"systems":[10],"share":[11],"the":[12,86,97,102,116,137],"same":[13],"benchmark":[14],"loader,":[15],"tool":[16],"access,":[17],"answer":[18],"contract,":[19],"usage":[20],"accounting,":[21],"and":[22,37,45,57,62,109,127],"trajectory":[23],"logging?":[24],"We":[25],"introduce":[26],"BenchAgent,":[27],"evaluation":[29],"framework":[30],"that":[31],"places":[32],"single-agent,":[33],"fixed":[34,143],"multi-agent":[35],"(MAS),":[36],"evolving":[38],"MAS":[39,84],"workflows":[40,52],"under":[41],"one":[42,80],"normalized":[43],"execution":[44],"logging":[46],"protocol.":[47],"BenchAgent":[48],"evaluates":[49],"these":[50],"substrate-internal":[51],"across":[53],"ten":[54],"reasoning,":[55],"coding,":[56],"tool-use":[58],"benchmarks":[59],"with":[60],"GPT-4.1,":[61],"separately":[63],"reports":[64],"a":[65,72,120,142],"Protocol-Aligned":[66],"External":[67],"(PAE)":[68],"GAIA":[69,118],"study":[70],"of":[71,81],"runtime-generated":[73],"workflow.":[74],"Under":[75],"SI":[76],"conditions,":[77],"at":[78],"most":[79],"six":[82],"tested":[83],"exceeds":[85],"matched":[87],"single-agent":[88],"anchor":[89],"on":[90,129],"benchmark-balanced":[91],"average":[92],"accuracy:":[93],"EvoAgent":[94],"lies":[95],"within":[96],"Wilson":[98],"one-run":[99],"guidance,":[100],"while":[101],"remaining":[103],"five":[104],"trail":[105],"by":[106],"2.56-11.29":[107],"points":[108,135],"occupy":[110],"expensive":[112],"accuracy-cost":[113],"trade-offs.":[114],"On":[115],"PAE":[117],"snapshot,":[119],"Claude-Code-style":[121],"runtime":[122],"reaches":[124],"66.72%":[125],"overall":[126],"69.23%":[128],"Level":[130],"3,":[131],"than":[133],"20":[134],"above":[136],"strongest":[138],"non-Claude":[139],"baseline,":[140],"Jarvis,":[141],"MAS.":[144]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-06T00:00:00"}