{"id":"https://openalex.org/W7164896397","doi":"https://doi.org/10.48550/arxiv.2606.15872","title":"SciOrch: Learning to Orchestrate Expert LLMs for Solving Frontier Multimodal Scientific Reasoning Tasks","display_name":"SciOrch: Learning to Orchestrate Expert LLMs for Solving Frontier Multimodal Scientific Reasoning Tasks","publication_year":2026,"publication_date":"2026-06-14","ids":{"openalex":"https://openalex.org/W7164896397","doi":"https://doi.org/10.48550/arxiv.2606.15872"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.15872","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.15872","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.2606.15872","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129752036","display_name":"Jingru Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Jingru","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120706712","display_name":"Xiangyuan Xue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Xiangyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138695861","display_name":"Lian Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Lian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138696371","display_name":"Wanghan Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Wanghan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136762731","display_name":"Siki Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Siki","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138742303","display_name":"Philip Torr","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Torr, Philip","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138729158","display_name":"Wanli Ouyang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ouyang, Wanli","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138732252","display_name":"Lei Bai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bai, Lei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5068281694","display_name":"Z. Yin","orcid":"https://orcid.org/0000-0003-4532-7544"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Zhenfei","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/T10028","display_name":"Topic Modeling","score":0.4934000074863434,"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/T10028","display_name":"Topic Modeling","score":0.4934000074863434,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.09070000052452087,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11948","display_name":"Machine Learning in Materials Science","score":0.06729999929666519,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/frontier","display_name":"Frontier","score":0.6604999899864197},{"id":"https://openalex.org/keywords/complementarity","display_name":"Complementarity (molecular biology)","score":0.5573999881744385},{"id":"https://openalex.org/keywords/orchestration","display_name":"Orchestration","score":0.478300005197525},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.4632999897003174},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.3862000107765198},{"id":"https://openalex.org/keywords/subject-matter-expert","display_name":"Subject-matter expert","score":0.33709999918937683},{"id":"https://openalex.org/keywords/efficient-frontier","display_name":"Efficient frontier","score":0.32199999690055847},{"id":"https://openalex.org/keywords/scientific-reasoning","display_name":"Scientific reasoning","score":0.31369999051094055}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6794000267982483},{"id":"https://openalex.org/C2778571376","wikidata":"https://www.wikidata.org/wiki/Q1355821","display_name":"Frontier","level":2,"score":0.6604999899864197},{"id":"https://openalex.org/C202269582","wikidata":"https://www.wikidata.org/wiki/Q2644277","display_name":"Complementarity (molecular biology)","level":2,"score":0.5573999881744385},{"id":"https://openalex.org/C199168358","wikidata":"https://www.wikidata.org/wiki/Q3367000","display_name":"Orchestration","level":3,"score":0.478300005197525},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.4632999897003174},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.4300000071525574},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.3862000107765198},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.35749998688697815},{"id":"https://openalex.org/C105002631","wikidata":"https://www.wikidata.org/wiki/Q4833645","display_name":"Subject-matter expert","level":3,"score":0.33709999918937683},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32760000228881836},{"id":"https://openalex.org/C51485801","wikidata":"https://www.wikidata.org/wiki/Q16966861","display_name":"Efficient frontier","level":3,"score":0.32199999690055847},{"id":"https://openalex.org/C2992562121","wikidata":"https://www.wikidata.org/wiki/Q3817808","display_name":"Scientific reasoning","level":2,"score":0.31369999051094055},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3098999857902527},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3012999892234802},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.30090001225471497},{"id":"https://openalex.org/C72161134","wikidata":"https://www.wikidata.org/wiki/Q5421219","display_name":"Expert elicitation","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2712000012397766},{"id":"https://openalex.org/C168065819","wikidata":"https://www.wikidata.org/wiki/Q845566","display_name":"Debugging","level":2,"score":0.26510000228881836},{"id":"https://openalex.org/C46681722","wikidata":"https://www.wikidata.org/wiki/Q4725589","display_name":"Alias","level":2,"score":0.2648000121116638},{"id":"https://openalex.org/C2776834041","wikidata":"https://www.wikidata.org/wiki/Q25346349","display_name":"Execution model","level":2,"score":0.26159998774528503},{"id":"https://openalex.org/C105152847","wikidata":"https://www.wikidata.org/wiki/Q5021685","display_name":"Call to action","level":2,"score":0.26080000400543213},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.26080000400543213},{"id":"https://openalex.org/C109168655","wikidata":"https://www.wikidata.org/wiki/Q242988","display_name":"Liberian dollar","level":2,"score":0.25870001316070557},{"id":"https://openalex.org/C2777548347","wikidata":"https://www.wikidata.org/wiki/Q5456937","display_name":"Flagging","level":2,"score":0.25619998574256897}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.15872","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.15872","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.2606.15872","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.15872","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":{"Frontier":[0],"scientific":[1,68],"reasoning":[2],"remains":[3],"a":[4,55,59,86,142],"major":[5],"challenge":[6],"for":[7,67],"large":[8],"language":[9],"models":[10,38,80],"(LLMs),":[11],"where":[12],"even":[13],"the":[14,49,136,158,166,175,187],"strongest":[15,159,167],"commercial":[16,79,161],"systems":[17],"fall":[18],"short":[19],"of":[20,190],"expert-level":[21],"performance.":[22],"A":[23],"closer":[24],"look":[25],"at":[26],"model":[27,47,62,162],"behavior":[28],"reveals":[29],"substantial":[30],"complementarity":[31],"that":[32,57,106],"single-model":[33],"evaluation":[34],"hides:":[35],"different":[36,41],"frontier":[37,65],"excel":[39],"on":[40,178],"question":[42],"types,":[43],"and":[44,84,113,134,148,165,181],"no":[45],"single":[46,160],"captures":[48],"full":[50],"picture.":[51],"We":[52,120],"present":[53],"SciOrch,":[54],"framework":[56],"trains":[58],"lightweight":[60],"8B":[61],"to":[63,77],"orchestrate":[64],"LLMs":[66],"reasoning.":[69],"The":[70],"orchestrator":[71,92,137],"decomposes":[72],"each":[73,100],"question,":[74],"delegates":[75],"sub-problems":[76],"selected":[78],"through":[81],"API":[82,104,188],"calls,":[83],"synthesizes":[85],"final":[87],"answer.":[88],"Training":[89],"such":[90],"an":[91,103],"is":[93,107],"fundamentally":[94],"harder":[95],"than":[96,185],"conventional":[97],"agentic":[98],"RL:":[99],"action":[101],"triggers":[102],"call":[105],"expensive":[108],"in":[109],"both":[110,179],"dollar":[111],"cost":[112,189],"latency,":[114],"making":[115],"standard":[116],"online":[117],"rollouts":[118],"infeasible.":[119],"address":[121],"this":[122],"with":[123,138,183],"MCTS-based":[124],"approach,":[125],"producing":[126],"diverse":[127],"orchestration":[128],"trajectories,":[129],"extracting":[130],"per-node":[131],"single-turn":[132],"samples,":[133],"optimizing":[135],"GRPO-style":[139],"training.":[140],"On":[141],"240-question":[143],"test":[144],"set":[145],"spanning":[146],"SGI-Reasoning":[147],"Scientists'":[149],"First":[150],"Exam,":[151],"SciOrch":[152],"reaches":[153],"56.66%":[154],"average":[155],"accuracy,":[156],"outperforming":[157],"by":[163,170],"3.74%":[164],"multi-agent":[168,192],"baseline":[169],"3.33%.":[171],"It":[172],"also":[173],"attains":[174],"best":[176],"accuracy":[177],"SGI":[180],"SFE":[182],"less":[184],"half":[186],"typical":[191],"methods.":[193]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-17T00:00:00"}