{"id":"https://openalex.org/W7224761853","doi":"https://doi.org/10.48550/arxiv.2610.39562","title":"Raincheck: Deferred Multimodal Prefill for Efficient Multi-Turn MLLM Inference","display_name":"Raincheck: Deferred Multimodal Prefill for Efficient Multi-Turn MLLM Inference","publication_year":2026,"publication_date":"2026-10-29","ids":{"openalex":"https://openalex.org/W7224761853","doi":"https://doi.org/10.48550/arxiv.2610.39562"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.39562","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.39562","pdf_url":"https://arxiv.org/pdf/2610.39562","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.39562","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5160217456","display_name":"Wenjie Huang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huang, Wenjie","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160195279","display_name":"Jun Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Jun","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129647450","display_name":"Hongyaoxing Gu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210165038","display_name":"University of Chinese Academy of Sciences","ror":"https://ror.org/05qbk4x57","country_code":"CN","type":"education","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Gu, Hongyaoxing","raw_affiliation_strings":["University of the Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of the Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160177050","display_name":"Haoyang Huang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huang, Haoyang","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160198637","display_name":"Tianqi Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xu, Tianqi","raw_affiliation_strings":["Carnegie Mellon University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041898142","display_name":"Xinyue Hao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I98677209","display_name":"University of Edinburgh","ror":"https://ror.org/01nrxwf90","country_code":"GB","type":"education","lineage":["https://openalex.org/I98677209"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Hao, Xinyue","raw_affiliation_strings":["University of Edinburgh"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Edinburgh","institution_ids":["https://openalex.org/I98677209"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160219554","display_name":"Xinyi Hu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hu, Xinyi","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076874267","display_name":"Xinle Du","orcid":"https://orcid.org/0000-0001-8918-9580","observed_orcids":["https://orcid.org/0000-0001-8918-9580"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Du, Xinle","raw_affiliation_strings":["Tsinghua University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tsinghua University","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160181997","display_name":"Jun Dai","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dai, Jun","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160225069","display_name":"Shuang Ge","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ge, Shuang","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068098522","display_name":"Lei Chen","orcid":"https://orcid.org/0000-0003-3081-0074","observed_orcids":["https://orcid.org/0000-0003-3081-0074"]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lei, Chen","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160226799","display_name":"Yue Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I84340266","display_name":"WHU \u2013 Otto Beisheim School of Management","ror":"https://ror.org/01yhahq71","country_code":"DE","type":"education","lineage":["https://openalex.org/I84340266"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Li, Yue","raw_affiliation_strings":["WHU -Otto Beisheim School of Management"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"WHU -Otto Beisheim School of Management","institution_ids":["https://openalex.org/I84340266"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5108223442","display_name":"Mingcheng Wan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wan, Mingcheng","raw_affiliation_strings":["Alibaba Group"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Group","institution_ids":["https://openalex.org/I45928872"]}]}],"institutions":[],"countries_distinct_count":4,"institutions_distinct_count":8,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.49799999594688416,"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"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.49799999594688416,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.2117999941110611,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.16220000386238098,"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/multimodal-large-language-models","display_name":"multimodal large language models","score":0.9649999737739563},{"id":"https://openalex.org/keywords/multi-turn-dialogue","display_name":"multi-turn dialogue","score":0.6890000104904175},{"id":"https://openalex.org/keywords/token-selection","display_name":"token selection","score":0.4429999887943268},{"id":"https://openalex.org/keywords/time-to-first-token","display_name":"time-to-first-token","score":0.8880000114440918},{"id":"https://openalex.org/keywords/qwen3-vl","display_name":"Qwen3-VL","score":0.6610000133514404},{"id":"https://openalex.org/keywords/qwen3-5","display_name":"Qwen3.5","score":0.6840000152587891}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8658999800682068},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7645000219345093},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6389999985694885},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5072000026702881},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.49630001187324524},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.49480000138282776},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.4668000042438507},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.445499986410141},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.4375999867916107},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39989998936653137},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.32519999146461487},{"id":"https://openalex.org/C123403432","wikidata":"https://www.wikidata.org/wiki/Q654068","display_name":"Visibility","level":2,"score":0.3122999966144562},{"id":"https://openalex.org/C92757383","wikidata":"https://www.wikidata.org/wiki/Q382497","display_name":"Affine transformation","level":2,"score":0.2994000017642975},{"id":"https://openalex.org/C94124525","wikidata":"https://www.wikidata.org/wiki/Q912550","display_name":"Categorization","level":2,"score":0.29679998755455017},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.28630000352859497},{"id":"https://openalex.org/C2988416141","wikidata":"https://www.wikidata.org/wiki/Q6031139","display_name":"Information loss","level":2,"score":0.2766000032424927},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.27559998631477356},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.2671000063419342},{"id":"https://openalex.org/C2780069185","wikidata":"https://www.wikidata.org/wiki/Q7977945","display_name":"Equivalence (formal languages)","level":2,"score":0.26570001244544983}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.39562","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.39562","pdf_url":"https://arxiv.org/pdf/2610.39562","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.39562","doi":"https://doi.org/10.48550/arxiv.2610.39562","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.39562","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.39562","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.39562","pdf_url":"https://arxiv.org/pdf/2610.39562","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7224761853.pdf","grobid_xml":"https://content.openalex.org/works/W7224761853.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7213328313","https://openalex.org/W4402857639","https://openalex.org/W4416036089","https://openalex.org/W7138030876","https://openalex.org/W7138268496","https://openalex.org/W4412888011","https://openalex.org/W4409363282","https://openalex.org/W4407208912","https://openalex.org/W7166821537","https://openalex.org/W7155097375"],"abstract_inverted_index":{"Multi-turn":[0],"conversations":[1],"with":[2,119,168],"multimodal":[3,13,43],"large":[4],"language":[5],"models":[6,126],"(MLLMs)":[7],"reuse":[8],"input":[9],"context,":[10],"but":[11,27],"long":[12],"inputs":[14],"increase":[15,172],"time":[16,175],"to":[17,142,162,165],"first":[18],"token":[19,48,178],"(TTFT).":[20],"Query-aware":[21],"hard":[22,144],"pruning":[23,56,129,145],"reduces":[24,110,156],"prefill":[25,44,80],"tokens":[26,59,69],"permanently":[28],"excludes":[29],"information":[30,52],"that":[31,45],"later":[32,51],"questions":[33],"may":[34],"require.":[35],"We":[36,114],"propose":[37],"Raincheck,":[38],"a":[39],"framework":[40],"for":[41,60],"deferred":[42],"decouples":[46],"initial":[47,61,121],"selection":[49],"from":[50],"availability.":[53],"An":[54],"existing":[55],"method":[57],"selects":[58],"prefill;":[62],"Raincheck":[63,131],"preserves":[64],"and":[65,92,104,117,127,136,148],"progressively":[66],"prefills":[67],"unselected":[68],"in":[70,102,173],"source":[71],"order":[72],"using":[73],"unused":[74],"compute":[75],"capacity":[76],"during":[77,85],"decode.":[78],"Unfinished":[79],"continues":[81],"after":[82,99],"answer":[83],"return":[84],"available":[86],"scheduling":[87],"gaps.":[88],"Attention":[89],"visibility":[90],"control":[91],"affine":[93],"state":[94],"mapping":[95],"enable":[96],"context":[97],"updates":[98],"generation":[100],"begins":[101],"full-attention":[103],"hybrid":[105],"recurrent":[106],"models.":[107],"Prefill-decode":[108],"fusion":[109],"repeated":[111],"weight":[112],"reads.":[113],"evaluate":[115],"Qwen3-VL-8B":[116],"Qwen3.5-9B":[118],"30%":[120],"multimodal-token":[122],"retention.":[123],"Across":[124],"both":[125],"four":[128],"methods,":[130],"recovers":[132],"on":[133,146],"average":[134],"77.7%":[135],"60.4%":[137],"of":[138],"the":[139,153],"quality":[140],"lost":[141],"matched":[143],"MultiVerse":[147],"ConvBench,":[149],"respectively.":[150],"On":[151],"MultiVerse,":[152],"SGLang":[154],"implementation":[155],"mean":[157,174],"first-turn":[158],"TTFT":[159],"by":[160],"up":[161],"24.1%":[163],"relative":[164],"full":[166],"prefill,":[167],"less":[169],"than":[170],"0.5%":[171],"per":[176],"output":[177],"(TPOT).":[179]},"counts_by_year":[],"updated_date":"2026-11-02T09:45:03.842371","created_date":"2026-10-31T00:00:00"}