AIMat/
Articles/
2609005135
  • Open Access
  • Article

Harness Engineered LLM Agents for Material Science: A Case Study on Perovskite Materials with Machine Learning and Materials Project

  • Lei Zhang 1,2,*,   
  • Tianle Wang 1

Received: 09 Jun 2026 | Revised: 17 Aug 2026 | Accepted: 04 Sep 2026 | Published: 11 Sep 2026

Abstract

The integration of large-language-model (LLM) agents into materials science requires a balance between adaptive reasoning and reliable scientific execution. Here, we present a Harness framework that converts agent proposals into structured and validated actions, connects them to the Materials Project and deterministic machine-learning tools, and records settings and computational evidence. The present perovskite band-gap case study uses an expert-defined oxide/halide partition as an initial subclass strategy; it does not claim autonomous discovery of an optimal material taxonomy. Fourteen regressors are compared under pooled and subclass-specific training using identical test entries across 50 matched stratified splits. Subclass-specific modeling directly improves several linear, nearest-neighbor, and ensemble regressors, but the effect is model dependent and is negative for SVR and LightGBM. A direct training-only K-means baseline produces a different partition; the chemical rule performs better for several regressors, whereas K-means performs better for LightGBM. In the current implementation, Harness provides the reproducible execution pathway for evaluating this scientific hypothesis. More broadly, the framework is designed to compare and refine candidate subclasses, descriptors, models, and physical-validation tools using downstream performance feedback. This work therefore provides a foundational demonstration of adaptive, tool-integrated workflow construction for materials science.

References 

  • 1.

    Zhang, L.; Liu, Z.; Ni, B.; et al. Large Language Models (LLMs) for Materials Design. Adv. Funct. Mater. 2026, 36, e25897. https://doi.org/10.1002/adfm.202525897.

  • 2.

    Ahlawat, D.; Mishra, V.; Singh, S.; et al. A Family of Large Language Models for Materials Research with Insights into Model Adaptability in Continued Pretraining. Nat. Mach. Intell. 2026, 8, 435–448. https://doi.org/10.1038/s42256-026-01199-8.

  • 3.

    Wang, X.; Zeng, Q.; Xu, D.-H.; et al. Accelerating Materials Discovery via AI-Agent Integration of Large Language Models and Simulation Tools. J. Mater. Informatics 2026, 6, 9. https://doi.org/10.20517/jmi.2025.69.

  • 4.

    Huang, L.; Zhou, Y.; Yang, J. Integrated Sharing Platform for Genetic Data of Rare and Precious Metal Materials. Comput. Mater. Contin. 2025, 85, 4587–4606. https://doi.org/10.32604/cmc.2025.068370.

  • 5.

    Vergara, D.; del Bosque, A.; Fernández-Arias, P. A Decade of Digital Twins in Materials Science and Engineering. Comput. Mater. Contin. 2025, 85, 41–64. https://doi.org/10.32604/cmc.2025.067881.

  • 6.

    Xu, P.; Ji, X.; Li, M.; et al. Virtual Sample Generation in Machine Learning Assisted Materials Design and Discovery. J. Mater. Informatics 2023, 3, 16. https://doi.org/10.20517/jmi.2023.18.

  • 7.

    Deng, H.; Wu, Y.; Wang, Y.; et al. PVmatAgent: A Large Language Model (LLM) Agent for Photovoltaic and Perovskite Material Design and Analysis. ChemRxiv 2026.

  • 8.

    Qiu, J.; Wang, Y.; Jiang, F.; et al. STALAgent: A Multi-Agent System Based on Large Language Model (LLM) for Steel and Alloy Design. ChemRxiv 2026.

  • 9.

    Zhang, L.; Zhou, J.; Chen, X. Data-Driven Exploration and First-Principles Analysis of Perovskite Material. J. Mater. Informatics 2024, 4, 13. https://doi.org/10.20517/jmi.2024.20.

  • 10.

    Xue, X.; Huang, X.; Wang, G. Materials Genome Engineering: A Promising Approach for the Development of High-Performance Metal–Organic Frameworks. Sci. Bull. 2022, 67, 1197–1200. https://doi.org/10.1016/j.scib.2022.05.003.

  • 11.

    Hegde, V.I.; Borg, C.K.H.; del Rosario, Z.; et al. Quantifying Uncertainty in High-Throughput Density Functional Theory: A Comparison of AFLOW, Materials Project, and OQMD. Phys. Rev. Mater. 2023, 7, 053805. https://doi.org/10.1103/PhysRevMaterials.7.053805.

  • 12.

    Jain, A.; Ong, S.P.; Hautier, G.; et al. Commentary: The Materials Project: A Materials Genome Approach to Accelerating Materials Innovation. APL Mater. 2013, 1, 011002. https://doi.org/10.1063/1.4812323.

  • 13.

    Park, Y.J.; Jerng, S.E.; Yoon, S.; et al. 1.5 Million Materials Narratives Generated by Chatbots. Sci. Data 2024, 11, 1060. https://doi.org/10.1038/s41597-024-03886-w.

  • 14.

    Kirklin, S.; Saal, J.E.; Meredig, B.; et al. The Open Quantum Materials Database (OQMD): Assessing the Accuracy of DFT Formation Energies. npj Comput. Mater. 2015, 1, 15010. https://doi.org/10.1038/npjcompumats.2015.10.

  • 15.

    Curtarolo, S.; Setyawan, W.; Hart, G.L.W.; et al. AFLOW: An Automatic Framework for High-Throughput Materials Discovery. Comput. Mater. Sci. 2012, 58, 218–226. https://doi.org/10.1016/j.commatsci.2012.02.005.

  • 16.

    Zhou, J.; Shen, L.; Costa, M.D.; et al. 2DMatPedia, an Open Computational Database of Two-Dimensional Materials from Top-down and Bottom-up Approaches. Sci. Data 2019, 6, 86. https://doi.org/10.1038/s41597-019-0097-3.

  • 17.

    Choudhary, K. The JARVIS Infrastructure Is All You Need for Materials Design. Comput. Mater. Sci. 2025, 259, 114063. https://doi.org/10.1016/j.commatsci.2025.114063.

  • 18.

    Zhou, Z.; Zhang, Y.; Wang, S.; et al. A Knowledge Graph Attention Network for the Cold-start Problem in Intelligent Manufacturing: Interpretability and Accuracy Improvement. Mater. Genome Eng. Adv. 2025, 3, e85. https://doi.org/10.1002/mgea.85.

  • 19.

    Mannan, S.; Bihani, V.; Krishnan, N.M.A.; et al. Navigating Energy Landscapes for Materials Discovery: Integrating Modeling, Simulation, and Machine Learning. Mater. Genome Eng. Adv. 2024, 2, e25. https://doi.org/10.1002/mgea.25.

  • 20.

    Chen, X.; He, S.; Paik, K.-W.; et al. Machine Learning-Driven Design and Optimization of Electronic Packaging: Applications and Future Developments. J. Mater. Informatics 2025, 5, 47. https://doi.org/10.20517/jmi.2025.26.

  • 21.

    Yang, Y.; Cao, L.; Ren, J.; et al. Artificial Intelligence for Fibrous Network Design and Mechanics. AI Sci. 2025, 1, 012001. https://doi.org/10.1088/3050-287X/adfaf7.

  • 22.

    Wang, W.Y.; Zhang, S.; Li, G.; et al. Artificial Intelligence Enabled Smart Design and Manufacturing of Advanced Materials: The Endless Frontier in AI + Era. Mater. Genome Eng. Adv. 2024, 2, e56. https://doi.org/10.1002/mgea.56.

  • 23.

    Zhang, L.; Deng, H. NJmat 2.0: User Instructions of Data-Driven Machine Learning Interface for Materials Science. Comput. Mater. Contin. 2025, 83, 1–11. https://doi.org/10.32604/cmc.2025.062666.

  • 24.

    He, M.; Zhang, L. Machine Learning and Symbolic Regression Investigation on Stability of MXene Materials. Comput. Mater. Sci. 2021, 196, 110578. https://doi.org/10.1016/j.commatsci.2021.110578.

  • 25.

    Butler, K.T.; Davies, D.W.; Cartwright, H.; et al. Machine Learning for Molecular and Materials Science. Nature 2018, 559, 547–555. https://doi.org/10.1038/s41586-018-0337-2.

  • 26.

    Muuga, K.; Knijff, L.; Zhang, C. Molecular Electrostatic Potentials from Machine Learning Models for Dipole and Quadrupole Predictions. AI Sci. 2026, 2, 025002. https://doi.org/10.1088/3050-287X/ae531a.

  • 27.

    Fang, Z.; Hsu, T.-W.; Yan, Q. Learning Atomic Representations for Data-Driven Materials Design. AI Sci. 2026, 2, 013001. https://doi.org/10.1088/3050-287X/ae45bf.

  • 28.

    Lin, X.; Wang, Y.; Zhang, Z.; et al. Scientific Claim Recognition via Staged Fine-Tuning with LoRA. Data Intell. 2025, 7, 303–335. https://doi.org/10.3724/2096-7004.di.2025.0009.

  • 29.

    Fu, F.; Li, Q.-Q.; Wang, F.; et al. Synergizing a Knowledge Graph and Large Language Model for Relay Catalysis Pathway Recommendation. Natl. Sci. Rev. 2025, 12, nwaf271. https://doi.org/10.1093/nsr/nwaf271.

  • 30.

    Zhang, C.; Lin, Q.; Zhu, B.; et al. SynAsk: Unleashing the Power of Large Language Models in Organic Synthesis. Chem. Sci. 2025, 16, 43–56. https://doi.org/10.1039/D4SC04757E.

  • 31.

    Soares, F.M.; Saraiva, A.M.; Pires, L.F.; et al. Exploring a Large Language Model for Transforming Taxonomic Data into OWL: Lessons Learned and Implications for Ontology Development. Data Intell. 2025, 7, 265–302. https://doi.org/10.3724/2096-7004.di.2025.0020.

  • 32.

    Xu, K.; Lu, Y.; Liu, L.; et al. MatterMind: A User-Friendly Material Design Platform Assisted by Large Language Model. Mater. Genome Eng. Adv. 2026, 4, e70075. https://doi.org/10.1002/mgea.70075.

  • 33.

    Yan, Z.; Liang, H.; Wang, J.; et al. PDGPT: A Large Language Model for Acquiring Phase Diagram Information in Magnesium Alloys. Mater. Genome Eng. Adv. 2024, 2, e77. https://doi.org/10.1002/mgea.77.

  • 34.

    Lv, S.; Peng, L.; Wu, W.; et al. Bridging Language Models and Computational Materials Science: A Prompt-driven Framework for Material Property Prediction. Mater. Genome Eng. Adv. 2025, 3, e70013. https://doi.org/10.1002/mgea.70013.

  • 35.

    Xu, Q.; Yu, Q.; Qian, L. Analysis and Evaluation of Open Source Scientific Entity Datasets. Data Intell. 2025, 7, 667–691. https://doi.org/10.3724/2096-7004.di.2024.0009.

  • 36.

    Yao, T.; Xu, A.; Ou, P.; et al. Let Materials Science Data Learn to Reason. AI Mater. 2026, 1, 4. https://doi.org/10.53941/aimat.2026.100004.

  • 37.

    He, J.; Guan, H.; Feng, W.; et al. Controlling Risks of AI in Chemical Science with Agents. AI Sci. 2025, 1, 015002. https://doi.org/10.1088/3050-287X/adfee5.

  • 38.

    Wang, W.; Rignanese, G.-M. AI for Science to Serve the AI-Driven Scientific Innovation. AI Sci. 2025, 1, 010201. https://doi.org/10.1088/3050-287X/add0cb.

  • 39.

    Chaudhari, A.; Ock, J.; Barati Farimani, A. Modular Large Language Model Agents for Multi-Task Computational Materials Science. Commun. Mater. 2026, 7, 131. https://doi.org/10.1038/s43246-025-00994-x.

  • 40.

    Bran, A.M.; Cox, S.; Schilter, O.; et al. Augmenting Large Language Models with Chemistry Tools. Nat. Mach. Intell. 2024, 6, 525–535. https://doi.org/10.1038/s42256-024-00832-8.

  • 41.

    Shi, Z.; Xin, C.; Huo, T.; et al. A Fine-Tuned Large Language Model Based Molecular Dynamics Agent for Code Generation to Obtain Material Thermodynamic Parameters. Sci. Rep. 2025, 15, 10295. https://doi.org/10.1038/s41598-025-92337-6.

  • 42.

    Wang, Z.; Huang, H.; Zhao, H.; et al. DREAMS: Density Functional Theory Based Research Engine for Agentic Materials Simulation. arXiv 2025, arXiv:2507.14267.

  • 43.

    Liu, J.; Zhu, T.; Ye, C.; et al. VASPilot: MCP-Facilitated Multi-Agent Intelligence for Autonomous VASP Simulations. arXiv 2025, arXiv:2508.07035.

  • 44.

    Zou, Y.; Cheng, A.H.; Aldossary, A.; et al. El Agente: An Autonomous Agent for Quantum Chemistry. Matter 2025, 8, 102263. https://doi.org/10.1016/j.matt.2025.102263.

  • 45.

    Nduma, R.; Park, H.; Walsh, A. Crystalyse: A Multi-Tool Agent for Materials Design. arXiv 2025, arXiv:2512.00977.

  • 46.

    Inizan, T.J.; Yang, S.; Kaplan, A.; et al. System of Agentic AI for the Discovery of Metal-Organic Frameworks. arXiv 2025, arXiv:2504.14110.

  • 47.

    Tian, J.; Sobczak, M.T.; Patil, D.; et al. CrossMatAgent: AI-Assisted Design of Manufacturable Metamaterial Patterns via Multi-Agent Generative Framework. Adv. Intell. Discov. 2025, 2, e202500063. https://doi.org/10.1002/aidi.202500063.

  • 48.

    Liu, G.; Yang, S.; Zhong, Y. Masgent: An AI-Assisted Materials Simulation Agent. arXiv 2025, arXiv:2512.23010.

  • 49.

    Gan, J.; Zhong, P.; Du, Y.; et al. MatLLMSearch: Crystal Structure Discovery with Evolution-Guided Large Language Models. arXiv 2025, arXiv:2502.20933.

  • 50.

    Yang, J.; Kobayashi, Y.; Demura, M. AI Agents for Automating Materials Research: A Case Study of Crystal Plasticity Simulations. Sci. Technol. Adv. Mater. Methods 2026, 6, 2630445. https://doi.org/10.1080/27660400.2026.2630445.

  • 51.

    Song, Z.; Lu, S.; Zhou, Q.; et al. T2MAT (Text-to-Material): A Universal Agent for Generating Material Structures with Goal Properties from a Single Sentence. Acta Physico-Chimica Sin. 2026, 42, 100213. https://doi.org/10.1016/j.actphy.2025.100213.

  • 52.

    Chiang, Y.; Hsieh, E.; Chou, C.-H.; et al. LLaMP: Large Language Model Made Powerful for High-Fidelity Materials Knowledge Retrieval. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing; Association for Computational Linguistics: Stroudsburg, PA, USA, 2025; pp. 25200–25232. https://doi.org/10.18653/v1/2025.emnlp-main.1280.

  • 53.

    Zhang, H.; Song, Y.; Hou, Z.; et al. HoneyComb: A Flexible LLM-Based Agent System for Materials Science. In Findings of the Association for Computational Linguistics: EMNLP 2024; Association for Computational Linguistics: Stroudsburg, PA, USA, 2024; pp 3369–3382. https://doi.org/10.18653/v1/2024.findings-emnlp.192.

  • 54.

    Ansari, M.; Moosavi, S.M. Agent-Based Learning of Materials Datasets from Scientific Literature. Digit. Discov. 2024, 3, 2607–2617. https://doi.org/10.1039/D4DD00252K.

  • 55.

    Jia, S.; Zhang, C.; Fung, V. LLMatDesign: Autonomous Materials Discovery with Large Language Models. arXiv 2024, arXiv:2406.13163.

  • 56.

    Ralph, J.; Martynov, P.; Ma, X.; et al. Identifier Service in the Mindat Database: Persistent and Structured Access to Massive Records of Minerals and Other Natural Materials. Data Intell. 2024, 7, 692–711. https://doi.org/10.3724/2096-7004.di.2024.0021.

  • 57.

    Hu, Z.; You, J.; Huang, X.; et al. Automatic Construction of Knowledge Graphs from Scientific Literature for Copper-Based Composites. Data Intell. 2024, 6, 1168–1189. https://doi.org/10.3724/2096-7004.di.2024.0038.

  • 58.

    Li, L.; Xu, K.; Su, R.; et al. Data Speaks: LLM-Enabled Evidence Auditing Resets the “Migration Barrier” Playbook for Solid-State Electrolytes. AI Mater. 2026, 1, 2. https://doi.org/10.53941/aimat.2026.100002.

  • 59.

    Yin, R.A.; Ren, Z.; Yin, Z.; et al. Collaborative AI Enhances Image Understanding in Materials Science. AI Mater. 2026, 1, 6. https://doi.org/10.53941/aimat.2026.100006.

  • 60.

    Zhong, H.; Zhu, S. AI Harness Engineering: A Runtime Substrate for Foundation-Model Software Agents. arXiv 2026, arXiv:2605.13357.

  • 61.

    Miret, S.; Krishnan, N.M.A. Enabling Large Language Models for Real-World Materials Discovery. Nat. Mach. Intell. 2025, 7, 991–998. https://doi.org/10.1038/s42256-025-01058-y.

  • 62.

    Ding, K.; Yu, J.; Huang, J.; et al. SciToolAgent: A Knowledge-Graph-Driven Scientific Agent for Multitool Integration. Nat. Comput. Sci. 2025, 5, 962–972. https://doi.org/10.1038/s43588-025-00849-y.

  • 63.

    Li, X.; Huang, Z.; Quan, S.; et al. SLM-MATRIX: A Multi-Agent Trajectory Reasoning and Verification Framework for Enhancing Language Models in Materials Data Extraction. npj Comput. Mater. 2025, 11, 241. https://doi.org/10.1038/s41524-025-01719-x.

  • 64.

    Soleymanibrojeni, M.; Aydin, R.; Guedes-Sobrinho, D.; et al. GENIUS: An Agentic AI Framework for Autonomous Design and Execution of Simulation Protocols. Commun. Mater. 2026, 7, 115. https://doi.org/10.1038/s43246-026-01167-0.

  • 65.

    Horton, M.K.; Huck, P.; Yang, R.X.; et al. Accelerated Data-Driven Materials Science with the Materials Project. Nat. Mater. 2025, 24, 1522–1532. https://doi.org/10.1038/s41563-025-02272-0.

  • 66.

    Batatia, I.; Benner, P.; Chiang, Y.; et al. A Foundation Model for Atomistic Materials Chemistry. J. Chem. Phys. 2025, 163, 184110. https://doi.org/10.1063/5.0297006.

  • 67.

    Zhang, L.; Lin, D.-Y.; Wang, H.; et al. Active Learning of Uniformly Accurate Interatomic Potentials for Materials Simulation. Phys. Rev. Mater. 2019, 3, 023804. https://doi.org/10.1103/PhysRevMaterials.3.023804.

  • 68.

    Kulichenko, M.; Barros, K.; Lubbers, N.; et al. Uncertainty-Driven Dynamics for Active Learning of Interatomic Potentials. Nat. Comput. Sci. 2023, 3, 230–239. https://doi.org/10.1038/s43588-023-00406-5.

  • 69.

    Yoo Willow, S.; Kim, D.G.; Sundheep, R.; et al. Active Sparse Bayesian Committee Machine Potential for Isothermal–Isobaric Molecular Dynamics Simulations. Phys. Chem. Chem. Phys. 2024, 26, 22073–22082. https://doi.org/10.1039/D4CP01801J.

  • 70.

    Liang, X.; Klarbring, J.; Baldwin, W.J.; et al. Structural Dynamics Descriptors for Metal Halide Perovskites. J. Phys. Chem. C 2023, 127, 19141–19151. https://doi.org/10.1021/acs.jpcc.3c03377.

  • 71.

    Liang, X.; Klarbring, J.; Walsh, A. Phase Stability and Transformations in Lead Mixed Halide Perovskites from Machine Learning Force Fields. Chem. Mater. 2025, 37, 9392–9405. https://doi.org/10.1021/acs.chemmater.5c01730.

  • 72.

    Kim, K.S. Machine Learning for Accelerating Energy Materials Discovery: Bridging Quantum Accuracy with Computational Efficiency. Adv. Energy Mater. 2026, 16, e03356. https://doi.org/10.1002/aenm.202503356.

Share this article:
Download PDF
Supplementary Materials
DOI
10.53941/aimat.2026.100013
Issue
Volume 1, Issue 1
How to Cite
Zhang, L.; Wang, T. Harness Engineered LLM Agents for Material Science: A Case Study on Perovskite Materials with Machine Learning and Materials Project. AI for Materials 2026, 1 (1), 13. https://doi.org/10.53941/aimat.2026.100013.
RIS
BibTex
Copyright & License
article copyright Image
Copyright (c) 2026 by the authors.
Article Metrics
50
Article Views
0
Citations