AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations
- 1Nanyang Technological University , Singapore
- 2Singapore Management University
Abstract
Density functional theory (DFT) serves as the basis for computational discovery in materials science and chemistry, yet each calculation demands extensive human effort: adjusting algorithms when convergence stalls, revising plans when unexpected physics emerges, and inserting steps as intermediate results reshape the problem. Existing LLM-based agents automate only the initial planning stage, producing a full execution plan upfront and leaving all subsequent adaptation to hand-crafted rules. As a result, these workflows remain fragile, do not generalize well beyond pre-planned scenarios, and often require expert intervention when failures or unexpected intermediate results require changes to the calculation path. Here, we introduce AutoDFT, a closed-loop multi-agent framework that embeds LLM reasoning into every stage of the DFT lifecycle, where a strategic planner produces a skeletal plan of step objectives; a step planner generates numerical parameters just in time from preceding results; and a monitor-recover-reflect cycle diagnoses failures, repairs them, and revises the plan when the evidence justifies it. We demonstrate both breadth and depth: breadth on VASPBench, a purpose-built benchmark spanning 34 tasks and 9 DFT calculation types, where AutoDFT achieves 94.1% task-level success with GPT-5.2; and depth on established materials databases, where AutoDFT produces quantitatively reliable property predictions across electronic, magnetic, and energetic properties. By closing the loop between planning and execution, AutoDFT enables experimentalists without deep computational expertise to obtain reliable first-principles results.
Related
- Can Autonomous LLM Agents Execute Multireference Quantum Chemistry Calculations?
- PyFMO: A Python Framework for Fragment Molecular Orbital Calculations
- Single-and Multi-bead Coarse-grained Form Factors for SAXS Calculations
- Biquaternion Algebra with Bilinear Multiplication: A General, Elegant, and Computationally Advantageous Framework for Relativistic Electronic Structure Calculations on CPUs and GPUs
- Autonomous bioisosteric replacement for multi-property optimization in drug design
BibTeX
@misc{tang2026autodft,
title = {{AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations}},
author = {Bijun Tang and Zhonghan Zhang and Yue Li and Xinrun Wag and Yanchen Deng and Yuhao Lu and Bijun Tang and Zheng Liu and Bo An},
howpublished = {arXiv (Cornell University)},
year = {2026},
month = may,
doi = {10.48550/arxiv.2605.26179},
url = {https://arxiv.org/abs/2605.26179},
note = {CC BY},
}
RIS
TY - UNPB
TI - AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations
AU - Bijun Tang
AU - Zhonghan Zhang
AU - Yue Li
AU - Xinrun Wag
AU - Yanchen Deng
AU - Yuhao Lu
AU - Bijun Tang
AU - Zheng Liu
AU - Bo An
PY - 2026
DA - 2026/05/25
PB - arXiv (Cornell University)
DO - 10.48550/arxiv.2605.26179
UR - https://arxiv.org/abs/2605.26179
AB - Density functional theory (DFT) serves as the basis for computational discovery in materials science and chemistry, yet each calculation demands extensive human effort: adjusting algorithms when convergence stalls, revising plans when unexpected physics emerges, and inserting steps as intermediate results reshape the problem. Existing LLM-based agents automate only the initial planning stage, producing a full execution plan upfront and leaving all subsequent adaptation to hand-crafted rules. As a result, these workflows remain fragile, do not generalize well beyond pre-planned scenarios, and often require expert intervention when failures or unexpected intermediate results require changes to the calculation path. Here, we introduce AutoDFT, a closed-loop multi-agent framework that embeds LLM reasoning into every stage of the DFT lifecycle, where a strategic planner produces a skeletal plan of step objectives; a step planner generates numerical parameters just in time from preceding results; and a monitor-recover-reflect cycle diagnoses failures, repairs them, and revises the plan when the evidence justifies it. We demonstrate both breadth and depth: breadth on VASPBench, a purpose-built benchmark spanning 34 tasks and 9 DFT calculation types, where AutoDFT achieves 94.1% task-level success with GPT-5.2; and depth on established materials databases, where AutoDFT produces quantitatively reliable property predictions across electronic, magnetic, and energetic properties. By closing the loop between planning and execution, AutoDFT enables experimentalists without deep computational expertise to obtain reliable first-principles results.
SN - 2331-8422
ER -