{"id":"https://openalex.org/W7225567372","doi":"https://doi.org/10.20944/preprints202611.0780.v1","title":"Benchmarking Long-Horizon Agent Context Compaction with CompactionBench","display_name":"Benchmarking Long-Horizon Agent Context Compaction with CompactionBench","publication_year":2026,"publication_date":"2026-11-09","ids":{"openalex":"https://openalex.org/W7225567372","doi":"https://doi.org/10.20944/preprints202611.0780.v1"},"language":"en","primary_location":{"id":"doi:10.20944/preprints202611.0780.v1","doi":"https://doi.org/10.20944/preprints202611.0780.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202611.0780.v1","pdf_url":null,"source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.20944/preprints202611.0780.v1","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5161284268","display_name":"Banruo Liu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Banruo Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037793114","display_name":"Haoran Qiu","orcid":"https://orcid.org/0000-0002-8071-1130","observed_orcids":["https://orcid.org/0000-0002-8071-1130"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haoran Qiu","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0002-8071-1130","affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090311560","display_name":"\u00cd\u00f1igo Goiri","orcid":"https://orcid.org/0000-0003-2591-4012","observed_orcids":["https://orcid.org/0000-0003-2591-4012"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"\u00cd\u00f1igo Goiri","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161272625","display_name":"Jingran Zhou","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jingran Zhou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161293885","display_name":"Rodrigo Fonseca","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rodrigo Fonseca","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089615986","display_name":"Ricardo Bianchini","orcid":"https://orcid.org/0000-0001-5971-5084","observed_orcids":["https://orcid.org/0000-0001-5971-5084"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ricardo Bianchini","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161302976","display_name":"Fan Lai","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan Lai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5045708568","display_name":"Esha Choukse","orcid":"https://orcid.org/0000-0003-0371-5522","observed_orcids":["https://orcid.org/0000-0003-0371-5522"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Esha Choukse","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5161284268"],"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/T10260","display_name":"Software Engineering Research","score":0.9926999807357788,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10260","display_name":"Software Engineering Research","score":0.9926999807357788,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10430","display_name":"Software Engineering Techniques and Practices","score":0.001500000013038516,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.0007999999797903001,"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/token-consumption","display_name":"token consumption","score":0.5590000152587891},{"id":"https://openalex.org/keywords/github-copilot","display_name":"GitHub Copilot","score":0.7799999713897705},{"id":"https://openalex.org/keywords/agent-harnesses","display_name":"agent harnesses","score":0.44699999690055847}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7383999824523926},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.6365000009536743},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5367000102996826},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.49939998984336853},{"id":"https://openalex.org/C196715460","wikidata":"https://www.wikidata.org/wiki/Q1414356","display_name":"Compaction","level":2,"score":0.4697999954223633},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.40230000019073486},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3840999901294708},{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.36500000953674316},{"id":"https://openalex.org/C177212765","wikidata":"https://www.wikidata.org/wiki/Q627335","display_name":"Workflow","level":2,"score":0.3465999960899353},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2994000017642975},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.28859999775886536},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.288100004196167},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.2856000065803528},{"id":"https://openalex.org/C170858558","wikidata":"https://www.wikidata.org/wiki/Q1394144","display_name":"Automatic summarization","level":2,"score":0.28049999475479126},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.27730000019073486},{"id":"https://openalex.org/C117447612","wikidata":"https://www.wikidata.org/wiki/Q1412670","display_name":"Software quality","level":4,"score":0.2694000005722046},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.2678999900817871},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.2554999887943268},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.2508000135421753},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.2500999867916107}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.20944/preprints202611.0780.v1","doi":"https://doi.org/10.20944/preprints202611.0780.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202611.0780.v1","pdf_url":null,"source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.20944/preprints202611.0780.v1","doi":"https://doi.org/10.20944/preprints202611.0780.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202611.0780.v1","pdf_url":null,"source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7223705248","https://openalex.org/W7223673979","https://openalex.org/W7222851099","https://openalex.org/W7214110372","https://openalex.org/W7220934534","https://openalex.org/W7215726049","https://openalex.org/W7162289651","https://openalex.org/W7223674602","https://openalex.org/W7222181240","https://openalex.org/W7223050698"],"abstract_inverted_index":{"Long-running":[0],"coding":[1,43,95],"agents":[2],"rely":[3],"on":[4,22],"context":[5,11],"compaction":[6,17,33,103,123,141],"to":[7,136,175],"overcome":[8],"finite":[9],"LLM":[10,50],"windows.":[12],"Existing":[13],"harnesses":[14],"typically":[15],"perform":[16],"synchronously,":[18],"placing":[19],"expensive":[20],"summarization":[21,156],"the":[23,131,144],"critical":[24,145],"path":[25,146],"and":[26,35,73,109,133,154,177],"disrupting":[27],"prefix-cache":[28,76],"reuse.":[29,77,150],"We":[30,116],"deep-dive":[31],"into":[32],"behavior":[34],"costs":[36],"by":[37,173,181],"analyzing":[38],"13.5M":[39],"GitHub":[40],"Copilot":[41],"agentic":[42],"sessions":[44,56],"from":[45],"3.2M":[46],"users,":[47],"comprising":[48],"760M":[49],"calls.":[51],"Although":[52],"only":[53],"7.8%":[54],"of":[55,63,101],"trigger":[57],"compaction,":[58],"they":[59],"account":[60],"for":[61],"44.2%":[62],"total":[64,171],"token":[65],"consumption.":[66],"Compaction":[67],"also":[68],"causes":[69],"substantial":[70],"foreground":[71],"delay":[72],"sharply":[74],"reduces":[75],"To":[78],"enable":[79],"systematic":[80],"evaluation,":[81],"we":[82],"first":[83],"introduce":[84],"CompactionBench,":[85],"a":[86,137,158],"benchmark":[87],"that":[88,125],"models":[89],"software":[90],"evolution":[91],"as":[92,157],"persistent,":[93],"multi-turn":[94],"sessions.":[96],"It":[97],"enables":[98],"controlled":[99],"evaluation":[100,108],"repeated":[102],"while":[104,147,169],"preserving":[105,148],"executable":[106],"task-level":[107],"reflecting":[110],"context-growth":[111],"patterns":[112],"observed":[113],"in":[114,130],"production.":[115],"then":[117],"present":[118],"APEC,":[119],"an":[120],"asynchronous":[121],"progressive":[122],"mechanism":[124],"prepares":[126],"fine-grained,":[127],"indexed":[128],"summaries":[129],"background":[132],"appends":[134],"them":[135],"stable":[138],"prefix,":[139],"moving":[140],"work":[142],"off":[143],"prefix":[149],"APEC":[151,165],"is":[152],"harness-agnostic":[153],"treats":[155],"black":[159],"box.":[160],"Evaluated":[161],"across":[162],"frontier":[163],"models,":[164],"improves":[166],"task":[167],"quality":[168],"reducing":[170],"cost":[172],"up":[174],"25.4%":[176],"agent":[178],"execution":[179],"time":[180],"9.6%.":[182]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-11-09T00:00:00"}