{"id":1189033,"date":"2026-10-07T13:55:02","date_gmt":"2026-10-07T20:55:02","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/agensh-scaling-organizational-intelligence-to-1024-agents\/"},"modified":"2026-10-09T15:47:48","modified_gmt":"2026-10-09T22:47:48","slug":"agensh-scaling-organizational-intelligence-to-1024-agents","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/agensh-scaling-organizational-intelligence-to-1024-agents\/","title":{"rendered":"Agensh: Scaling Organizational Intelligence to 1,024 Agents"},"content":{"rendered":"\n\n\n<p class=\"wp-block-paragraph\">A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support multi-agent systems. However, the scalability of current multi-agent harnesses is often constrained by a central orchestrator&#8217;s capacity to allocate tasks and coordinate workers. To address this limitation, we introduce Agensh, a scalable self-organized multi-agent harness without a central orchestrator: concurrent workers execute a multi-agent cooperation loop, continuously gathering context, claiming and self-assigning sub-tasks, taking action and sharing findings, verifying results, and merging progress in an asynchronous manner. The loop is supported by the agentic organization infrastructure comprising three components: a shared workspace holds proposed, ongoing, and completed work; a message interface lets workers communicate; and shared context retains reusable findings and work intentions. To test the scalability of Agensh, we evaluate it on the five hardest ProgramBench tasks with GPT-5.6-sol (high). Scaling from 1 to 128 agents raises the mean final test-pass rate from 19.31% to 28.78%, an approximately 49% relative improvement. Larger organizations reach comparable test-pass rates earlier. On pandoc, scaling from 1 to 1,024 agents raises the final test-pass rate from 33.89% to 55.06%. Worker trajectories further show that different forms of self-organized cooperation gradually emerges and standardizes as the organization grows. These results reveal the number of agents as a new scaling dimension for multi-agent organizations to expand the frontier of general intelligence, offering a practical solution for complex tasks under hard latency constraints or time budgets.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support multi-agent systems. However, the scalability of current multi-agent harnesses is often constrained by a central orchestrator&#8217;s capacity to allocate tasks and coordinate workers. To address this limitation, we introduce Agensh, a scalable self-organized multi-agent harness without a [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"text","value":"Zhi-Hao Zhan","user_id":0},{"type":"user_nicename","value":"Ting Song","user_id":"34357"},{"type":"user_nicename","value":"Li Dong","user_id":"38811"},{"type":"text","value":"Shao-Han Huang","user_id":0},{"type":"text","value":"Jian-Xun Lian","user_id":0},{"type":"user_nicename","value":"Yan Xia","user_id":"34972"},{"type":"user_nicename","value":"Furu Wei","user_id":"31830"}],"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"arXiv","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"","msr_doi":"","msr_arxiv_id":"2609.26781","msr_mag_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_release_tracker_id":"","msr_highlight_type":"","msr_date_display_format":"","msr_main_download_label":"","msr_external_link_label":"","msr_doi_label":"","msr_published_date":"2026-09-22","msr_startdate":"","msr_presentation_date":"","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_year":2026,"msr_month":9,"msr_day":22,"msr_microsoftintellectualproperty":false,"msr_pub_id":"619c97976973f096a6410acbf62b441e0d5a1cd5","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.26781","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"619c97976973f096a6410acbf62b441e0d5a1cd5","msr_s2_pdf_url":"","msr_citation_count_updated":"","msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":32,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[{"provider":"s2","id":"619c97976973f096a6410acbf62b441e0d5a1cd5"},{"provider":"arxiv","id":"2609.26781"},{"provider":"corpusid","id":"292226972"}],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556],"msr-publication-type":[270373],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[263203,246691],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1189033","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-locale-en_us","msr-field-of-study-computation-and-language","msr-field-of-study-computer-science"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2026-09-22","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"arXiv","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.26781","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"619c97976973f096a6410acbf62b441e0d5a1cd5","msr_influential_citations":0,"msr_reference_count":32,"msr_arxiv_id":"2609.26781","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Zhi-Hao Zhan","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Ting Song","user_id":34357,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Ting Song"},{"type":"user_nicename","value":"Li Dong","user_id":38811,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Li Dong"},{"type":"text","value":"Shao-Han Huang","user_id":0,"rest_url":false},{"type":"text","value":"Jian-Xun Lian","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Yan Xia","user_id":34972,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Yan Xia"},{"type":"user_nicename","value":"Furu Wei","user_id":31830,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Furu Wei"}],"msr_impact_theme":[],"msr_research_lab":[199560,1012650],"msr_event":[],"msr_group":[],"msr_project":[],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"misc","related_content":[],"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1189033","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":4,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1189033\/revisions"}],"predecessor-version":[{"id":1189517,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1189033\/revisions\/1189517"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1189033"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=1189033"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1189033"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1189033"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1189033"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1189033"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1189033"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1189033"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1189033"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=1189033"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1189033"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1189033"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1189033"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1189033"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}