Agensh: Scaling Organizational Intelligence to 1,024 Agents

Hugging Face Daily Papers Papers
multi-agent-system scalability self-organization ai-research cooperation latency-optimization

Summary

Agensh is a scalable self-organized multi-agent system without a central orchestrator that improves performance on complex tasks by scaling the number of agents, showing significant test-pass rate increases on benchmarks like ProgramBench and pandoc.

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'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.
Original Article
View Cached Full Text

Cached at: 09/23/26, 03:34 PM

Paper page - Agensh: Scaling Organizational Intelligence to 1,024 Agents

Source: https://huggingface.co/papers/2609.26781

Abstract

Amulti-agentsystemcanreducelatencyoncomplextasksbyexecutingworkconcurrently.Severalpioneeringharnessframeworkssupportmulti-agentsystems.However,thescalabilityofcurrentmulti-agentharnessesisoftenconstrainedbyacentralorchestrator’scapacitytoallocatetasksandcoordinateworkers.Toaddressthislimitation,weintroduceAgensh,ascalableself-organizedmulti-agentharnesswithoutacentralorchestrator:concurrentworkersexecuteamulti-agentcooperationloop,continuouslygatheringcontext,claimingandself-assigningsub-tasks,takingactionandsharingfindings,verifyingresults,andmergingprogressinanasynchronousmanner.Theloopissupportedbytheagenticorganizationinfrastructurecomprisingthreecomponents:asharedworkspaceholdsproposed,ongoing,andcompletedwork;amessageinterfaceletsworkerscommunicate;andsharedcontextretainsreusablefindingsandworkintentions.TotestthescalabilityofAgensh,weevaluateitonthefivehardestProgramBenchtaskswithGPT-5.6-sol(high).Scalingfrom1to128agentsraisesthemeanfinaltest-passratefrom19.31%to28.78%,anapproximately49%relativeimprovement.Largerorganizationsreachcomparabletest-passratesearlier.Onpandoc,scalingfrom1to1,024agentsraisesthefinaltest-passratefrom33.89%to55.06%.Workertrajectoriesfurthershowthatdifferentformsofself-organizedcooperationgraduallyemergesandstandardizesastheorganizationgrows.Theseresultsrevealthenumberofagentsasanewscalingdimensionformulti-agentorganizationstoexpandthefrontierofgeneralintelligence,offeringapracticalsolutionforcomplextasksunderhardlatencyconstraintsortimebudgets.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2609\.26781

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.26781 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.26781 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.26781 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Apodex 1.1: Scaling Agentic Intelligence for Complex Work

Hugging Face Daily Papers

Apodex 1.1 improves sustained, verifiable progress on complex real-world tasks by scaling executable environments and training agents for long-horizon coordination, achieving leading performance with a smaller 35B-parameter model.

@hyperagent: 42 agents. 216 threads. One dashboard. Every agent gets its own prompt, tools, skills, and budget. Deploy specialized a…

X AI KOLs Timeline

Hyperagent launched a multi-agent system that deploys specialized AI agents with individual prompts, tools, and budgets to autonomously perform complex business tasks such as recruiting, analytics, and marketing campaigns. The platform features a unified dashboard, self-improvement capabilities, and deep integration with existing business tools.

Position: Agentic AI System Is a Foreseeable Pathway to AGI

arXiv cs.AI

This paper argues that monolithic scaling of a single model is insufficient for achieving AGI and proposes Agentic AI with multi-agent collaboration as a necessary paradigm, demonstrating theoretically that agentic systems achieve exponentially superior generalization and sample efficiency.