AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations

2026-05-25

AI Summary

This work addresses the fragility and limited generalizability of conventional density functional theory (DFT) workflows, which often rely on manual intervention and fail to handle runtime convergence issues or anomalies. To overcome these limitations, the authors propose the first closed-loop multi-agent framework that integrates large language models (e.g., GPT-5.2) throughout the entire DFT lifecycle. The system employs a strategic planner, a step-level planner, and a monitor–recover–reflect loop to enable dynamic adaptation and autonomous decision-making, moving beyond rigid, pre-scripted protocols by revising computational pathways in response to intermediate results. Evaluated on the VASPBench benchmark—comprising 34 tasks across nine calculation types—the method achieves a 94.1% success rate and demonstrates accurate prediction of electronic, magnetic, and energetic properties in real materials databases, significantly enhancing robustness and accessibility for non-experts seeking reliable first-principles results.

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.

Problem

• Density Functional Theory
• Autonomous Calculation
• Adaptive Planning
• Computational Materials Science
• LLM-based Agents

Innovation

• closed-loop multi-agent
• autonomous DFT
• LLM reasoning
• adaptive planning
• computational materials science