@inproceedings{dong-etal-2026-coembod,
title = "{C}o{E}mbod: A Long-Horizon Benchmark for Multi-Agent Embodied Cooperation with Diverse Inter-Agent Dependencies",
author = "Dong, Xuan and
Han, Zhe and
Yu, Ru and
Niu, Tianhao and
Cai, Muzhen and
lv, Xinda and
Zhu, Qingfu and
Zhang, Weinan and
Che, Wanxiang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2247/",
pages = "51112--51131",
ISBN = "979-8-89176-459-0",
abstract = "Embodied planning is dedicated to the goal of building agents capable of perceiving, reasoning, and executing long-horizon tasks in physical environments. However, existing embodied benchmarks predominantly focus on single-agent execution or restricted multi-agent settings, leaving long-horizon tasks that involve complex inter-agent dependencies largely underexplored despite their prevalence in real-world embodied scenarios. To address this gap, we introduce \textbf{CoEmbod}, the first realistic 3D simulated embodied benchmark for long-horizon multi-agent embodied cooperation with diverse inter-agent dependencies. CoEmbod contains 155 expert-curated tasks across 40 scenes and 10845 objects, covering three cooperation regimes: spatial distribution, parallel constraint, and failure recovery. These regimes cover a broad spectrum of cooperative requirements, ranging from distributed task allocation to synchronized execution and assisted failure recovery. CoEmbod further provides a process-level evaluation framework with Budgeted Cooperative Success (BCS), which jointly measures goal completion, cooperation fidelity, and execution efficiency. Experiments reveal that even the leading LMM powering a well-established agentic workflow achieves only 20.23{\%} success, highlighting the difficulty of long-horizon multi-agent cooperation with diverse inter-agent dependencies in complex embodied environments."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="dong-etal-2026-coembod">
<titleInfo>
<title>CoEmbod: A Long-Horizon Benchmark for Multi-Agent Embodied Cooperation with Diverse Inter-Agent Dependencies</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xuan</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhe</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ru</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianhao</namePart>
<namePart type="family">Niu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Muzhen</namePart>
<namePart type="family">Cai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinda</namePart>
<namePart type="family">lv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingfu</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weinan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Embodied planning is dedicated to the goal of building agents capable of perceiving, reasoning, and executing long-horizon tasks in physical environments. However, existing embodied benchmarks predominantly focus on single-agent execution or restricted multi-agent settings, leaving long-horizon tasks that involve complex inter-agent dependencies largely underexplored despite their prevalence in real-world embodied scenarios. To address this gap, we introduce CoEmbod, the first realistic 3D simulated embodied benchmark for long-horizon multi-agent embodied cooperation with diverse inter-agent dependencies. CoEmbod contains 155 expert-curated tasks across 40 scenes and 10845 objects, covering three cooperation regimes: spatial distribution, parallel constraint, and failure recovery. These regimes cover a broad spectrum of cooperative requirements, ranging from distributed task allocation to synchronized execution and assisted failure recovery. CoEmbod further provides a process-level evaluation framework with Budgeted Cooperative Success (BCS), which jointly measures goal completion, cooperation fidelity, and execution efficiency. Experiments reveal that even the leading LMM powering a well-established agentic workflow achieves only 20.23% success, highlighting the difficulty of long-horizon multi-agent cooperation with diverse inter-agent dependencies in complex embodied environments.</abstract>
<identifier type="citekey">dong-etal-2026-coembod</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2247/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>51112</start>
<end>51131</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CoEmbod: A Long-Horizon Benchmark for Multi-Agent Embodied Cooperation with Diverse Inter-Agent Dependencies
%A Dong, Xuan
%A Han, Zhe
%A Yu, Ru
%A Niu, Tianhao
%A Cai, Muzhen
%A lv, Xinda
%A Zhu, Qingfu
%A Zhang, Weinan
%A Che, Wanxiang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F dong-etal-2026-coembod
%X Embodied planning is dedicated to the goal of building agents capable of perceiving, reasoning, and executing long-horizon tasks in physical environments. However, existing embodied benchmarks predominantly focus on single-agent execution or restricted multi-agent settings, leaving long-horizon tasks that involve complex inter-agent dependencies largely underexplored despite their prevalence in real-world embodied scenarios. To address this gap, we introduce CoEmbod, the first realistic 3D simulated embodied benchmark for long-horizon multi-agent embodied cooperation with diverse inter-agent dependencies. CoEmbod contains 155 expert-curated tasks across 40 scenes and 10845 objects, covering three cooperation regimes: spatial distribution, parallel constraint, and failure recovery. These regimes cover a broad spectrum of cooperative requirements, ranging from distributed task allocation to synchronized execution and assisted failure recovery. CoEmbod further provides a process-level evaluation framework with Budgeted Cooperative Success (BCS), which jointly measures goal completion, cooperation fidelity, and execution efficiency. Experiments reveal that even the leading LMM powering a well-established agentic workflow achieves only 20.23% success, highlighting the difficulty of long-horizon multi-agent cooperation with diverse inter-agent dependencies in complex embodied environments.
%U https://aclanthology.org/2026.emnlp-main.2247/
%P 51112-51131
Markdown (Informal)
[CoEmbod: A Long-Horizon Benchmark for Multi-Agent Embodied Cooperation with Diverse Inter-Agent Dependencies](https://aclanthology.org/2026.emnlp-main.2247/) (Dong et al., EMNLP 2026)
ACL
- Xuan Dong, Zhe Han, Ru Yu, Tianhao Niu, Muzhen Cai, Xinda lv, Qingfu Zhu, Weinan Zhang, and Wanxiang Che. 2026. CoEmbod: A Long-Horizon Benchmark for Multi-Agent Embodied Cooperation with Diverse Inter-Agent Dependencies. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 51112–51131, Budapest, Hungary. Association for Computational Linguistics.