Can Autonomous LLM Agents Execute Multireference Quantum Chemistry Calculations?
Abstract
Multireference electronic-structure calculations remain difficult to automate because critical workflow decisions, including active-space selection, state averaging, convergence recovery, and state identification, traditionally rely on expert judgment. Here, we investigate whether an autonomous large language model (LLM) agent can perform these tasks without human intervention. The agent selects active spaces using literature-grounded analogies or explicitly documented chemical reasoning, generates and submits ORCA calculations, analyzes outputs, and records all decisions in an auditable reasoning log. Benchmarking against 558 vertical transition energies (VTEs) from QUESTDB shows that an unguided baseline agent achieves 24.9% coverage with a mean absolute error (MAE) of 0.373 eV. Introducing a structured decision ladder increases coverage to 44.1% while reducing the MAE to 0.339 eV. The largest gains are observed for double and Rydberg excitations, demonstrating that expert-informed procedural guidance substantially improves active-space construction and state identification. When provided with complete workflow information, the agent successfully reproduces published QUEST calculations with an MAE of only 23 meV, resolving 75% of target configurations within seven attempts. These results demonstrate that contemporary LLM agents can autonomously execute and reproduce complex multireference quantum-chemical workflows, while highlighting the importance of structured reasoning frameworks for achieving reliable high-throughput and high-fidelity electronic-structure calculations.
Related
- AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations
- Chemistry-infused bandit framework for interpretable autonomous campaign of electrocatalysis
- Exploitation of Complex Abelian Point Groups in Quantum-Chemical Calculations
- A tutorial on quantum scattering theory and coupled-channel calculations of bimolecular collisions
- Mitigating Errors in Variational Quantum Eigensolver Calculations through Extended Random Phase Approximation
BibTeX
@misc{lee2026autonomous,
title = {{Can Autonomous LLM Agents Execute Multireference Quantum Chemistry Calculations?}},
author = {Victor Chang Lee and James M. Rondinelli},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.13357},
eprint = {2609.13357},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.13357},
}
RIS
TY - UNPB
TI - Can Autonomous LLM Agents Execute Multireference Quantum Chemistry Calculations?
AU - Victor Chang Lee
AU - James M. Rondinelli
PY - 2026
DA - 2026/09/11
PB - arXiv
DO - 10.48550/arxiv.2609.13357
UR - https://arxiv.org/abs/2609.13357
AB - Multireference electronic-structure calculations remain difficult to automate because critical workflow decisions, including active-space selection, state averaging, convergence recovery, and state identification, traditionally rely on expert judgment. Here, we investigate whether an autonomous large language model (LLM) agent can perform these tasks without human intervention. The agent selects active spaces using literature-grounded analogies or explicitly documented chemical reasoning, generates and submits ORCA calculations, analyzes outputs, and records all decisions in an auditable reasoning log. Benchmarking against 558 vertical transition energies (VTEs) from QUESTDB shows that an unguided baseline agent achieves 24.9% coverage with a mean absolute error (MAE) of 0.373 eV. Introducing a structured decision ladder increases coverage to 44.1% while reducing the MAE to 0.339 eV. The largest gains are observed for double and Rydberg excitations, demonstrating that expert-informed procedural guidance substantially improves active-space construction and state identification. When provided with complete workflow information, the agent successfully reproduces published QUEST calculations with an MAE of only 23 meV, resolving 75% of target configurations within seven attempts. These results demonstrate that contemporary LLM agents can autonomously execute and reproduce complex multireference quantum-chemical workflows, while highlighting the importance of structured reasoning frameworks for achieving reliable high-throughput and high-fidelity electronic-structure calculations.
ER -