{"id":"https://openalex.org/W7203600535","doi":"https://doi.org/10.48550/arxiv.2608.14312","title":"Envs-FORGE: Frontier-Optimized Reward-Grounded Environment Synthesis for Agent RL","display_name":"Envs-FORGE: Frontier-Optimized Reward-Grounded Environment Synthesis for Agent RL","publication_year":2026,"publication_date":"2026-08-14","ids":{"openalex":"https://openalex.org/W7203600535","doi":"https://doi.org/10.48550/arxiv.2608.14312"},"language":null,"primary_location":{"id":"pmh:oai:arXiv.org:2608.14312","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.14312","pdf_url":"https://arxiv.org/pdf/2608.14312","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.14312","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137013765","display_name":"Xiaojun Wu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405261251","display_name":"The Hong Kong University of Science and Technology (Guangzhou)","ror":"https://ror.org/050h0vm43","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405261251"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wu, Xiaojun","raw_affiliation_strings":["IDEA Research","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IDEA Research","institution_ids":[]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I4405261251"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113257046","display_name":"Cehao Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405261251","display_name":"The Hong Kong University of Science and Technology (Guangzhou)","ror":"https://ror.org/050h0vm43","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405261251"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Cehao","raw_affiliation_strings":["IDEA Research","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IDEA Research","institution_ids":[]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I4405261251"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133568000","display_name":"Honghao Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405261251","display_name":"The Hong Kong University of Science and Technology (Guangzhou)","ror":"https://ror.org/050h0vm43","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405261251"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liu, Honghao","raw_affiliation_strings":["IDEA Research","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IDEA Research","institution_ids":[]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I4405261251"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136457894","display_name":"Xueyuan Lin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405261251","display_name":"The Hong Kong University of Science and Technology (Guangzhou)","ror":"https://ror.org/050h0vm43","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405261251"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lin, Xueyuan","raw_affiliation_strings":["The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I4405261251"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138270181","display_name":"Zhichao Shi","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Zhichao","raw_affiliation_strings":["DataArcTech Ltd ","IDEA Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"DataArcTech Ltd ","institution_ids":[]},{"raw_affiliation_string":"IDEA Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140007136","display_name":"Hao Zhou","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Hao","raw_affiliation_strings":["DataArcTech Ltd ","IDEA Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"DataArcTech Ltd ","institution_ids":[]},{"raw_affiliation_string":"IDEA Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122818996","display_name":"Xuhui Jiang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Xuhui","raw_affiliation_strings":["DataArcTech Ltd ","IDEA Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"DataArcTech Ltd ","institution_ids":[]},{"raw_affiliation_string":"IDEA Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136500675","display_name":"Chengjin Xu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Chengjin","raw_affiliation_strings":["DataArcTech Ltd ","IDEA Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"DataArcTech Ltd ","institution_ids":[]},{"raw_affiliation_string":"IDEA Research","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100405697","display_name":"Jia Li","orcid":"https://orcid.org/0000-0002-6362-4385","observed_orcids":["https://orcid.org/0000-0002-6362-4385","https://orcid.org/0000-0002-4346-8696"]},"institutions":[{"id":"https://openalex.org/I4405261251","display_name":"The Hong Kong University of Science and Technology (Guangzhou)","ror":"https://ror.org/050h0vm43","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405261251"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Jia","raw_affiliation_strings":["The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I4405261251"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5136038723","display_name":"Jian Guo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Jian","raw_affiliation_strings":["IDEA Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IDEA Research","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.2199999988079071,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.2199999988079071,"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/T10260","display_name":"Software Engineering Research","score":0.2151000052690506,"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/T10743","display_name":"Software Testing and Debugging Techniques","score":0.14550000429153442,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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/prompt-engineering","display_name":"prompt engineering","score":0.3880000114440918},{"id":"https://openalex.org/keywords/mixed-integer-linear-programming","display_name":"mixed-integer linear programming","score":0.7870000004768372},{"id":"https://openalex.org/keywords/test-case-generation","display_name":"test case generation","score":0.40400001406669617},{"id":"https://openalex.org/keywords/synthetic-training-data","display_name":"synthetic training data","score":0.40700000524520874},{"id":"https://openalex.org/keywords/swe-bench","display_name":"SWE-bench","score":0.640999972820282}],"concepts":[{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.8532000184059143},{"id":"https://openalex.org/C55166926","wikidata":"https://www.wikidata.org/wiki/Q2892946","display_name":"Oracle","level":2,"score":0.7077999711036682},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6837000250816345},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.5141000151634216},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.46000000834465027},{"id":"https://openalex.org/C154690210","wikidata":"https://www.wikidata.org/wiki/Q1668499","display_name":"Rewriting","level":2,"score":0.4390999972820282},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.42100000381469727},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.41350001096725464},{"id":"https://openalex.org/C2780821815","wikidata":"https://www.wikidata.org/wiki/Q5340806","display_name":"Portfolio","level":2,"score":0.4041999876499176},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.39640000462532043},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.36010000109672546},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3544999957084656},{"id":"https://openalex.org/C2779664074","wikidata":"https://www.wikidata.org/wiki/Q3518405","display_name":"Terminal (telecommunication)","level":2,"score":0.299699991941452},{"id":"https://openalex.org/C41045048","wikidata":"https://www.wikidata.org/wiki/Q202843","display_name":"Linear programming","level":2,"score":0.2824999988079071},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.2809000015258789},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.2799000144004822},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.27790001034736633},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.27459999918937683},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.25699999928474426},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.25690001249313354},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.14312","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.14312","pdf_url":"https://arxiv.org/pdf/2608.14312","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.14312","doi":"https://doi.org/10.48550/arxiv.2608.14312","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.14312","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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.14312","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.14312","pdf_url":"https://arxiv.org/pdf/2608.14312","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7203600535.pdf","grobid_xml":"https://content.openalex.org/works/W7203600535.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1,73],"(RL)":[2],"for":[3,122,169],"terminal":[4],"agents":[5],"needs":[6],"executable":[7],"training":[8],"environments":[9,188],"with":[10],"reliable":[11],"rewards":[12,56],"and":[13,22,75,103,143,158,171,189,203],"useful":[14],"difficulty.":[15],"Fixed":[16],"recipes":[17],"such":[18],"as":[19],"few-shot,":[20],"Self-Instruct,":[21],"Evol-Instruct":[23],"apply":[24],"the":[25,34,85,97,152,178,195,198,204],"same":[26,199,205],"prompting":[27,51],"policy":[28,36,52],"to":[29,83,141,149],"every":[30],"seed,":[31],"even":[32],"when":[33],"current":[35],"would":[37],"benefit":[38],"from":[39],"a":[40,50,71,77],"harder,":[41],"easier,":[42],"or":[43],"simply":[44],"different":[45],"task.":[46],"We":[47],"present":[48],"Envs-FORGE,":[49],"that":[53,87],"converts":[54],"verifier":[55],"into":[57],"per-seed":[58,78],"environment-synthesis":[59],"actions.":[60],"Envs-FORGE":[61,129],"estimates":[62],"seed":[63],"pass":[64],"rates,":[65],"scores":[66],"six":[67],"projection--direction":[68],"actions":[69],"around":[70],"target":[72],"frontier,":[74],"solves":[76],"mixed-integer":[79],"linear":[80],"program":[81],"(MILP)":[82],"choose":[84],"action":[86,92],"conditions":[88],"generation.":[89],"The":[90,112,208],"selected":[91],"drives":[93],"synchronized":[94],"rewriting":[95],"of":[96],"instruction,":[98],"fixtures,":[99],"oracle":[100],"solution,":[101],"tests,":[102],"Docker":[104],"environment;":[105],"only":[106],"gold-verified":[107],"bundles":[108],"enter":[109],"RL":[110],"training.":[111],"indexed":[113],"MILP":[114],"form":[115],"also":[116],"supports":[117],"optional":[118],"soft":[119],"skill":[120],"coverage":[121],"portfolio":[123],"planning.":[124],"On":[125],"Qwen":[126],"3.5":[127],"35B,":[128],"improves":[130,172],"Pass@1":[131],"over":[132],"Base":[133],"by":[134,156,174],"9.2":[135],"percentage":[136],"points":[137,145,176],"on":[138,146,164],"tb-core":[139,173],"(40.0%":[140],"49.2%)":[142],"6.4":[144],"tb-2.0":[147],"(23.0%":[148],"29.4%),":[150],"exceeding":[151],"strongest":[153],"fixed-recipe":[154],"baseline":[155],"2.4":[157],"2.1":[159],"points.":[160],"It":[161],"reaches":[162],"77.1%":[163],"SWE-bench":[165],"Verified":[166],"versus":[167],"73.4%":[168],"Base,":[170],"6.8--9.2":[175],"across":[177],"evaluated":[179],"4B--35B":[180],"models.":[181],"All":[182],"synthesis":[183,192],"methods":[184],"export":[185],"100":[186],"verified":[187],"use":[190],"2.27M--2.88M":[191],"tokens,":[193],"placing":[194],"comparison":[196],"at":[197,213],"downstream":[200],"training-set":[201],"size":[202],"operational":[206],"scale.":[207],"source":[209],"code":[210],"is":[211],"available":[212],"https://github.com/DataArcTech/DataArc-SynData-Toolkit/.":[214]},"counts_by_year":[],"updated_date":"2026-10-14T16:39:01.820549","created_date":"2026-08-18T00:00:00"}