{"id":"https://openalex.org/W7162079317","doi":"https://doi.org/10.1145/3786335.3813124","title":"XGrammar-2: Dynamic and Efficient Structured Generation Engine for Agentic LLMs","display_name":"XGrammar-2: Dynamic and Efficient Structured Generation Engine for Agentic LLMs","publication_year":2026,"publication_date":"2026-05-22","ids":{"openalex":"https://openalex.org/W7162079317","doi":"https://doi.org/10.1145/3786335.3813124"},"language":null,"primary_location":{"id":"doi:10.1145/3786335.3813124","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786335.3813124","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM Conference on AI and Agentic Systems","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3786335.3813124","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5008873501","display_name":"Linzhang Li","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Linzhang Li","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0007-7003-4037","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122439125","display_name":"Yixin Dong","orcid":null},"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":"Yixin Dong","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, USA"],"raw_orcid":"https://orcid.org/0009-0004-4553-5938","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048301958","display_name":"G Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guanjie Wang","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0003-4340-7572","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002695591","display_name":"Z Xu","orcid":null},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ziyi Xu","raw_affiliation_strings":["Shanghai Jiao Tong University, Shanghai, China"],"raw_orcid":"https://orcid.org/0009-0000-4411-9773","affiliations":[{"raw_affiliation_string":"Shanghai Jiao Tong University, Shanghai, China","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122275627","display_name":"Alexander Jiang","orcid":null},"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":"Alexander Jiang","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, USA"],"raw_orcid":"https://orcid.org/0009-0003-3090-5943","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, USA","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5136758310","display_name":"Tianqi Chen","orcid":"https://orcid.org/0000-0002-5744-3940"},"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":"Tianqi Chen","raw_affiliation_strings":["Carnegie Mellon University, Pittsburgh, USA and NVIDIA, Santa Clara, USA"],"raw_orcid":"https://orcid.org/0000-0002-5744-3940","affiliations":[{"raw_affiliation_string":"Carnegie Mellon University, Pittsburgh, USA and NVIDIA, Santa Clara, USA","institution_ids":["https://openalex.org/I74973139"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"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":"1009","last_page":"1022"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10126","display_name":"Logic, programming, and type systems","score":0.09839999675750732,"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/T10126","display_name":"Logic, programming, and type systems","score":0.09839999675750732,"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":0.08250000327825546,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.07109999656677246,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/reuse","display_name":"Reuse","score":0.7192000150680542},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.6848999857902527},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.6761000156402588},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.6021000146865845},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.508400022983551},{"id":"https://openalex.org/keywords/state","display_name":"State (computer science)","score":0.4404999911785126}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.732200026512146},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.7192000150680542},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.6848999857902527},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.6761000156402588},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.6021000146865845},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.508400022983551},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.4404999911785126},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.39579999446868896},{"id":"https://openalex.org/C174839445","wikidata":"https://www.wikidata.org/wiki/Q1134386","display_name":"Lock (firearm)","level":2,"score":0.30390000343322754},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.28780001401901245},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C2779662365","wikidata":"https://www.wikidata.org/wiki/Q5416694","display_name":"Event (particle physics)","level":2,"score":0.25189998745918274}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3786335.3813124","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786335.3813124","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM Conference on AI and Agentic Systems","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2601.04426","is_oa":true,"landing_page_url":"https://arxiv.org/abs/2601.04426","pdf_url":"https://arxiv.org/pdf/2601.04426","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.1145/3786335.3813124","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786335.3813124","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM Conference on AI and Agentic Systems","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W2099964107","https://openalex.org/W2124479173","https://openalex.org/W2988290054","https://openalex.org/W3003374142","https://openalex.org/W3214600982","https://openalex.org/W4387321091","https://openalex.org/W4387835442"],"related_works":[],"abstract_inverted_index":{"Modern":[0],"LLM":[1,129],"agents":[2],"increasingly":[3],"rely":[4],"on":[5,54],"dynamic":[6,47,78],"structured":[7,18,43,119],"generation,":[8],"such":[9],"as":[10],"tool":[11],"calling":[12],"and":[13,29,64,81,103,122],"response":[14],"protocols.":[15],"Unlike":[16],"traditional":[17],"generation":[19,44,120],"with":[20,69,94],"static":[21],"structures,":[22],"these":[23],"workloads":[24],"vary":[25],"both":[26],"across":[27,67,88],"requests":[28,68],"within":[30],"a":[31,42],"request,":[32],"posing":[33],"new":[34],"challenges":[35],"to":[36],"existing":[37],"engines.":[38],"We":[39],"present":[40],"XGrammar-2,":[41],"engine":[45],"for":[46,60,77,84],"agentic":[48],"workloads.":[49],"Our":[50],"design":[51],"is":[52],"based":[53],"two":[55],"key":[56],"ideas:":[57],"first-class":[58],"support":[59],"tag-triggered":[61],"structure":[62],"switching,":[63],"fine-grained":[65],"reuse":[66,87],"different":[70],"output":[71],"structures.":[72],"Concretely,":[73],"XGrammar-2":[74,110],"introduces":[75],"TagDispatch":[76],"structural":[79],"dispatching":[80],"Cross-Grammar":[82],"Cache":[83],"substructure-level":[85],"cache":[86],"grammars.":[89],"It":[90],"further":[91],"improves":[92],"efficiency":[93],"an":[95],"Earley-based":[96],"adaptive":[97],"token":[98],"mask":[99],"cache,":[100],"just-in-time":[101],"compilation,":[102],"repetition":[104],"state":[105],"compression.":[106],"Experiments":[107],"show":[108],"that":[109],"achieves":[111],"over":[112],"6":[113],"\u00d7":[114],"faster":[115],"compilation":[116],"than":[117],"prior":[118],"engines,":[121],"incurs":[123],"near-zero":[124],"end-to-end":[125],"overhead":[126],"in":[127],"modern":[128],"serving":[130],"systems.":[131]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-05-23T00:00:00"}