{"id":"https://openalex.org/W7155050529","doi":"https://doi.org/10.48550/arxiv.2604.17148","title":"Graph-of-Agents: A Graph-based Framework for Multi-Agent LLM Collaboration","display_name":"Graph-of-Agents: A Graph-based Framework for Multi-Agent LLM Collaboration","publication_year":2026,"publication_date":"2026-04-18","ids":{"openalex":"https://openalex.org/W7155050529","doi":"https://doi.org/10.48550/arxiv.2604.17148"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.17148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.17148","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.2604.17148","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134098945","display_name":"Sukwon Yun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yun, Sukwon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134135595","display_name":"Jie Peng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Jie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134184835","display_name":"Pingzhi Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Pingzhi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134189317","display_name":"Wendong Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Wendong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134105192","display_name":"Jie Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134126922","display_name":"James Zou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, James","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134183835","display_name":"Guohao Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Guohao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134153332","display_name":"Tianlong Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Tianlong","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/T11273","display_name":"Advanced Graph Neural Networks","score":0.5378000140190125,"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"}},"topics":[{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.5378000140190125,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.1137000024318695,"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/T10028","display_name":"Topic Modeling","score":0.08730000257492065,"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/pooling","display_name":"Pooling","score":0.7195000052452087},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6168000102043152},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5386999845504761},{"id":"https://openalex.org/keywords/relevance","display_name":"Relevance (law)","score":0.5209000110626221},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.4327000081539154},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.430400013923645}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7842000126838684},{"id":"https://openalex.org/C70437156","wikidata":"https://www.wikidata.org/wiki/Q7228652","display_name":"Pooling","level":2,"score":0.7195000052452087},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6168000102043152},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5386999845504761},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.5209000110626221},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.4327000081539154},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.430400013923645},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.38499999046325684},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.37540000677108765},{"id":"https://openalex.org/C854659","wikidata":"https://www.wikidata.org/wiki/Q1859284","display_name":"Message passing","level":2,"score":0.3361999988555908},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3303000032901764},{"id":"https://openalex.org/C62611344","wikidata":"https://www.wikidata.org/wiki/Q1062658","display_name":"Node (physics)","level":2,"score":0.31310001015663147},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.26980000734329224},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.25690001249313354},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.25380000472068787},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2522999942302704}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.17148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.17148","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.2604.17148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.17148","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":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"With":[0],"an":[1,184],"ever-growing":[2,240],"zoo":[3],"of":[4,38,142,187,238],"LLMs":[5,189],"and":[6,48,92,177,223],"benchmarks,":[7],"the":[8,77,100,139,143,148,236,239],"need":[9],"to":[10,29,110,124,128,137,161],"orchestrate":[11],"multiple":[12,191],"models":[13],"for":[14,64,234],"improved":[15],"task":[16,90],"performance":[17,197],"has":[18],"never":[19],"been":[20],"more":[21,144],"pressing.":[22],"While":[23],"frameworks":[24],"like":[25],"Mixture-of-Agents":[26],"(MoA)":[27],"attempt":[28],"coordinate":[30],"LLMs,":[31],"they":[32],"often":[33],"fall":[34],"short":[35],"in":[36],"terms":[37],"(1)":[39],"selecting":[40,75],"relevant":[41,79,122,126,145],"agents,":[42,202],"(2)":[43],"facilitating":[44],"effective":[45],"intra-agent":[46],"communication,":[47],"(3)":[49],"integrating":[50],"responses":[51,106,141,150],"efficiently.":[52],"In":[53],"this":[54],"work,":[55],"we":[56,96],"propose":[57],"Graph-of-Agents":[58],"(GoA),":[59],"a":[60,163,216,231],"new":[61],"graph-based":[62,154],"framework":[63],"modeling":[65],"multi-agent":[66,205],"LLM":[67,206,241],"communication.":[68],"Our":[69],"approach":[70],"begins":[71],"with":[72,183],"node":[73],"sampling,":[74],"only":[76,199],"most":[78],"agents":[80,102,123,212],"by":[81,103,133],"leveraging":[82],"model":[83],"cards":[84],"that":[85,208],"summarize":[86],"each":[87],"model's":[88],"domain,":[89],"specialization,":[91],"other":[93],"characteristics.":[94],"Next,":[95],"construct":[97],"edges":[98],"between":[99],"selected":[101,201],"evaluating":[104],"their":[105,130],"against":[107],"one":[108],"another":[109],"determine":[111],"relevance":[112],"ordering.":[113],"Directed":[114],"message":[115,135,227],"passing":[116,136],"is":[117,244],"then":[118],"performed":[119],"from":[120],"highly":[121],"less":[125],"ones":[127],"enhance":[129],"responses,":[131],"followed":[132],"reverse":[134],"refine":[138],"original":[140],"agents.":[146],"Finally,":[147],"updated":[149],"are":[151],"aggregated":[152],"via":[153],"pooling":[155],"(e.g.,":[156],"max":[157],"or":[158],"mean":[159],"pooling)":[160],"produce":[162],"single,":[164],"unified":[165],"answer.":[166],"We":[167],"evaluate":[168],"GoA":[169,194,219],"on":[170],"diverse":[171],"multi-domain":[172],"benchmarks":[173,179],"(MMLU,":[174],"MMLU-Pro,":[175],"GPQA)":[176],"domain-specific":[178],"(MATH,":[180],"HumanEval,":[181],"MedMCQA),":[182],"agent":[185],"pool":[186],"6":[188,211],"spanning":[190],"domains.":[192],"Surprisingly,":[193],"achieves":[195],"superior":[196],"using":[198],"3":[200],"outperforming":[203],"recent":[204],"baselines":[207],"utilize":[209],"all":[210],"simultaneously.":[213],"By":[214],"adopting":[215],"graph":[217],"structure,":[218],"offers":[220],"both":[221],"scalability":[222],"effectiveness":[224],"through":[225],"structured":[226],"passing-positioning":[228],"it":[229],"as":[230],"strong":[232],"candidate":[233],"navigating":[235],"challenges":[237],"zoo.":[242],"Code":[243],"available":[245],"at:":[246],"https://github.com/UNITES-Lab/GoA.":[247]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-22T00:00:00"}