{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,6]],"date-time":"2026-07-06T05:14:18Z","timestamp":1783314858616,"version":"3.54.6"},"reference-count":25,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,6]],"date-time":"2026-07-06T00:00:00Z","timestamp":1783296000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Robot. AI"],"abstract":"<jats:p>Mars rover missions require decision-support systems that can interpret terrain, telemetry, environmental conditions, and mission objectives under delayed communication with Earth. This study evaluates whether multi-agent orchestration improves simulated Mars rover decision support compared with a single-agent baseline. A controlled benchmark of 100 synthetic mission-inspired rover scenarios was evaluated using OpenAI GPT-4o and GPT-5.5, with five repeated runs per scenario and architecture. Model-facing scenario inputs were separated from evaluator-side labels so that expected actions and hazards were reserved for scoring only. Performance was measured using decision accuracy, exact and substring-based semantic hazard F1, hazard error counts, latency, token usage, scenario-level paired statistical comparisons, and GPT-4o specialist-agent ablations. Across the tested OpenAI configurations, the single-agent architecture showed numerical advantages in decision accuracy and hazard-label alignment, but these decision-quality differences were not consistently significant under scenario-level statistical analysis with Holm-Bonferroni adjustment. The only decision-quality metric remaining significant was GPT-5.5 exact hazard F1, although absolute values were very low. The most reliable difference was computational efficiency: the single-agent architecture required substantially lower latency and token usage than the prompt-defined multi-agent orchestration architecture. Multi-agent orchestration generated broader hazard lists, including plausible non-canonical observations, but did not reliably improve aggregate decision accuracy or hazard F1. These findings suggest that, for short-context, tool-less, static decision-support tasks where all relevant context is available in a single input, multi-agent orchestration should be treated as a cost-bearing design choice rather than an assumed improvement. The study contributes a reproducible architecture-level benchmark for evaluating when LLM-based orchestration is worth its operational cost in mission-inspired workflows.<\/jats:p>","DOI":"10.3389\/frobt.2026.1877762","type":"journal-article","created":{"date-parts":[[2026,7,6]],"date-time":"2026-07-06T04:20:27Z","timestamp":1783311627000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["OpenAI single-agent LLM architecture reduces computational overhead relative to multi-agent orchestration in a simulated mars rover decision-support benchmark"],"prefix":"10.3389","volume":"13","author":[{"given":"Dan","family":"Sanabria","sequence":"first","affiliation":[{"name":"Type 3 Innovations, LLC","place":["Alexandria, VA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,6]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/aero50100.2021.9438337","article-title":"Machine learning based path planning for improved rover navigation","volume":"50100","author":"Abcouwer","year":"2021","journal-title":"IEEE Aerosp. Conf."},{"key":"B2","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1109\/taes.2022.3169126","article-title":"Essential technologies and concepts for massive space exploration: challenges and opportunities","volume":"59","author":"Arzo","year":"2023","journal-title":"IEEE Trans. Aerosp. Electron. Syst."},{"key":"B3","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/HPEC58863.2023.10363561","article-title":"Advanced ultra low-power deep learning applications with neuromorphic computing","volume-title":"2023 IEEE High Performance Extreme Computing Conference (HPEC)","author":"Barnell","year":"2023"},{"key":"B4","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/aero.2015.7119105","article-title":"SHERLOC: scanning habitable environments with raman and luminescence for organics and chemicals","author":"Beegle","year":"2015","journal-title":"IEEE Aerosp. Conf."},{"key":"B5","article-title":"On the opportunities and risks of foundation models","author":"Bommasani","year":"2021","journal-title":"arXiv Preprint arXiv:2108.07258"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/aero50100.2021.9438289","article-title":"Autonomous safe landing site detection for a future Mars science helicopter","volume":"50100","author":"Brockers","year":"2021","journal-title":"IEEE Aerosp. Conf."},{"key":"B7","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"B9","doi-asserted-by":"crossref","DOI":"10.1109\/ICoEIT63558.2025.11211513","article-title":"Design and implementation of multi-agent systems: a lang Graph-based approach","volume-title":"2025 International Conference on Engineering Innovations and Technologies (Icoeit)","author":"Das","year":"2025"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.31031\/aes.2021.2.000540","article-title":"A study on centralized and decentralized swarm robotics architecture for part delivery system","volume":"2","author":"Dimakos","year":"2021","journal-title":"Acad. J. Eng. Stud."},{"key":"B11","doi-asserted-by":"crossref","DOI":"10.1109\/CCNC65079.2026.11366321","article-title":"Goal-oriented multi-agent reinforcement learning for decentralized agent teams","volume-title":"2026 IEEE 23rd Consumer Communications and Networking Conference (CCNC)","author":"Du","year":"2026"},{"key":"B12","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/ccaaw.2017.8001880","article-title":"Multi-objective reinforcement learning-based deep neural networks for cognitive space communications","author":"Ferreira","year":"2017","journal-title":"2017 Cognitive Commun. Aerosp. Appl. Workshop (CCAA)"},{"key":"B13","unstructured":"Generating structured outputs from language models: benchmark and studies\n          \n          \n            \n              Geng\n              S.\n            \n            \n              Cooper\n              H.\n            \n            \n              Moskal\n              M.\n            \n            \n              Jenkins\n              S.\n            \n            \n              Berman\n              J.\n            \n            \n              Ranchin\n              N.\n            \n          \n          \n          2025"},{"key":"B14","unstructured":"MLAgentBench: evaluating language agents on machine learning experimentation\n          \n          \n            \n              Huang\n              Q.\n            \n            \n              Vora\n              J.\n            \n            \n              Liang\n              P.\n            \n            \n              Leskovec\n              J.\n            \n          \n          \n          2023"},{"key":"B15","unstructured":"AgentBench: evaluating LLMs as agents\n          \n          \n            \n              Liu\n              X.\n            \n            \n              Yu\n              H.\n            \n            \n              Zhang\n              H.\n            \n            \n              Xu\n              Y.\n            \n            \n              Lei\n              X.\n            \n            \n              Lai\n              H.\n            \n          \n          \n          2023"},{"key":"B16","doi-asserted-by":"crossref","DOI":"10.1109\/ICDMW69685.2025.00406","article-title":"Can reasoning LLMs eliminate conformity in multi-agent systems?","volume-title":"2025 IEEE International Conference on Data Mining Workshops (ICDMW)","author":"Liu","year":"2025"},{"key":"B17","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1007\/s11214-020-00765-9","article-title":"The Mars 2020 engineering cameras and microphone on the perseverance rover: a next-generation imaging system for Mars exploration","volume":"216","author":"Maki","year":"2020","journal-title":"Space Sci. Rev."},{"key":"B18","article-title":"Techedsat-13: the first flight of a neuromorphic processor","author":"Murbach","year":"2022","journal-title":"NASA Tech. Rep. Serv. (NTRS)"},{"key":"B19","doi-asserted-by":"publisher","first-page":"29","DOI":"10.52202\/078357-0005","article-title":"Swarm robotics: a new paradigm in robotic space exploration","author":"Omar","year":"2024","journal-title":"IAF Space Explor. Symp."},{"key":"B20","unstructured":"ToolLLM: facilitating large language models to master 16000+ real-world APIs\n          \n          \n            \n              Qin\n              Y.\n            \n            \n              Liang\n              S.\n            \n            \n              Ye\n              Y.\n            \n            \n              Zhu\n              K.\n            \n            \n              Yan\n              L.\n            \n            \n              Lu\n              Y.\n            \n          \n          \n          2023"},{"key":"B21","doi-asserted-by":"publisher","first-page":"2123","DOI":"10.1109\/lra.2020.2970650","article-title":"Where to map? Iterative Rover-copter path planning for Mars exploration","volume":"5","author":"Sasaki","year":"2020","journal-title":"IEEE Robotics Automation Lett."},{"key":"B22","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"B23","doi-asserted-by":"publisher","first-page":"eadi3099","DOI":"10.1126\/scirobotics.adi3099","article-title":"Autonomous robotics is driving perseverance rover\u2019s progress on Mars","volume":"8","author":"Verma","year":"2023","journal-title":"Sci. Robotics"},{"key":"B24","doi-asserted-by":"publisher","first-page":"5373","DOI":"10.1109\/igarss46834.2022.9884096","article-title":"Machine learning on Mars: open challenges, similarities and differences to Earth remote sensing","author":"Wilhelm","year":"2022","journal-title":"IGARSS 2022 - 2022 IEEE Int. Geoscience Remote Sens. Symposium"},{"key":"B25","unstructured":"AutoGen: enabling next-gen LLM applications via multi-agent conversation\n          \n          \n            \n              Wu\n              Q.\n            \n            \n              Bansal\n              G.\n            \n            \n              Zhang\n              J.\n            \n            \n              Wu\n              Y.\n            \n            \n              Li\n              B.\n            \n            \n              Zhu\n              E.\n            \n          \n          \n          2023"},{"key":"B26","article-title":"ReAct: synergizing reasoning and acting in language models","author":"Yao","year":"2022","journal-title":"arXiv preprint arXiv:2210.03629"}],"container-title":["Frontiers in Robotics and AI"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2026.1877762\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,6]],"date-time":"2026-07-06T04:20:32Z","timestamp":1783311632000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2026.1877762\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,6]]},"references-count":25,"alternative-id":["10.3389\/frobt.2026.1877762"],"URL":"https:\/\/doi.org\/10.3389\/frobt.2026.1877762","relation":{},"ISSN":["2296-9144"],"issn-type":[{"value":"2296-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,6]]},"article-number":"1877762"}}