arxiv: 2602.04850 · v2 · submitted 2026-02-04 · ⚛️ physics.chem-ph · cs.AI· cs.MA

Recognition: 1 theorem link

· Lean Theorem

El Agente Quntur: A research collaborator agent for quantum chemistry

Juan B. P\'erez-S\'anchez , Yunheng Zou , Jorge A. Campos-Gonzalez-Angulo , Marcel M\"uller , Ignacio Gustin , Andrew Wang , Han Hao , Tsz Wai Ko

show 7 more authors

Changhyeok Choi Eric S. Isbrandt Mohammad Ghazi Vakili Hanyong Xu Chris Crebolder Varinia Bernales Al\'an Aspuru-Guzik

Authors on Pith no claims yet

Pith reviewed 2026-05-16 06:45 UTC · model grok-4.3

classification ⚛️ physics.chem-ph cs.AIcs.MA

keywords quantum chemistryORCAmulti-agent systemsAI agentscomputational chemistryresearch automationin silico experiments

0 comments

The pith

A hierarchical multi-agent AI reasons over literature and documentation to run and analyze any ORCA 6.0 quantum chemistry calculation

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces El Agente Quntur as a hierarchical multi-agent system that serves as a research collaborator for quantum chemistry. It replaces fixed procedural rules with reasoning over software documentation and scientific literature to plan, execute, adapt, and interpret experiments. The design uses general composable actions and guided integration of quantum-chemical concepts with program internals. This setup supports every calculation type in ORCA 6.0 and aims to lower the barrier for non-expert users. The work also maps current limits of such agents and sketches a path to more autonomous systems.

Core claim

Quntur supports the full range of calculations available in ORCA 6.0 and reasons over software documentation and scientific literature to plan, execute, adapt, and analyze in silico chemistry experiments following best practices.

What carries the argument

Hierarchical multi-agent architecture that performs reasoning-driven decisions via general composable actions and guided deep research linking quantum-chemical concepts to software syntax and logic.

Load-bearing premise

The agent's reasoning over abstract quantum-chemical concepts and software internals will consistently produce correct, best-practice decisions without human oversight or frequent errors.

What would settle it

Running Quntur on a set of standard benchmark molecules and finding that it repeatedly selects methods, basis sets, or analysis steps that contradict established literature recommendations for those systems.

read the original abstract

Quantum chemistry is a foundational enabling tool for the fields of chemistry, materials science, computational biology and others. Despite of its power, the practical application of quantum chemistry simulations remains in the hands of qualified experts due to methodological complexity, software heterogeneity, and the need for informed interpretation of results. To bridge the accessibility gap for these tools and expand their reach to chemists with broader backgrounds, we introduce El Agente Quntur, a hierarchical, multi-agent AI system designed to operate not merely as an automation tool but as a research collaborator for computational quantum chemistry. Quntur was designed following three main strategies: i) elimination of hard-coded procedural policies in favour of reasoning-driven decisions, ii) construction of general and composable actions that facilitate generalization and efficiency, and iii) implementation of guided deep research to integrate abstract quantum-chemical reasoning across subdisciplines and a detailed understanding of the software's internal logic and syntax. Although instantiated in ORCA, these design principles are applicable to research agents more generally and easily expandable to additional quantum chemistry packages and beyond. Quntur supports the full range of calculations available in ORCA 6.0 and reasons over software documentation and scientific literature to plan, execute, adapt, and analyze in silico chemistry experiments following best practices. We discuss the advances and current bottlenecks in agentic systems operating at the research level in computational chemistry, and outline a roadmap toward a fully autonomous end-to-end computational chemistry research agent.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Quntur paper outlines a promising multi-agent design for quantum chemistry but provides no validation data on its reasoning performance.

read the letter

The main point with this paper is that it describes a hierarchical multi-agent system called El Agente Quntur designed to act as a collaborator for quantum chemistry work with ORCA, but it offers no benchmarks or test results to show how well the reasoning actually performs. What stands out as new is the architecture built around three strategies: replacing hard-coded policies with reasoning-driven decisions, using general composable actions to improve flexibility, and incorporating guided deep research that pulls from both quantum chemical literature and the software's own documentation. This setup is meant to handle planning, execution, adaptation, and analysis of calculations following best practices across all of ORCA 6.0's features, and the authors indicate the same principles could extend to other packages. The paper does a solid job explaining these design choices in plain terms and making the case for why this approach could reduce the need for expert oversight in running simulations. It positions the agent as more than an automation script by emphasizing integration of abstract concepts with practical software logic. The main soft spot is the complete absence of quantitative evaluation. There are no success rates on sample tasks, no error analysis for issues like invalid syntax or suboptimal method choices, and no ablation studies on the literature retrieval component. This leaves the key assumption about consistent correct reasoning untested and makes it difficult to assess real-world utility. This kind of paper is aimed at researchers developing AI tools for computational science, especially those interested in agent-based systems for chemistry. Someone looking for ideas on multi-agent designs for research tasks could find the strategies worth considering. It deserves a serious referee because the implementation details are concrete enough for experts to evaluate the architecture and suggest concrete improvements, particularly around adding validation experiments. I recommend sending it for peer review to get constructive feedback on the design and to push the authors toward including performance data in any revision.

Referee Report

2 major / 1 minor

Summary. The paper introduces El Agente Quntur, a hierarchical multi-agent AI system designed to serve as a research collaborator for quantum chemistry. It operates on ORCA 6.0, using reasoning over software documentation and scientific literature to plan, execute, adapt, and analyze computational experiments according to best practices. The design avoids hard-coded policies in favor of reasoning-driven decisions, employs composable actions, and implements guided deep research, with the principles intended to be generalizable to other software packages.

Significance. If the claims regarding reliable autonomous reasoning hold, this work has the potential to significantly advance the field by making sophisticated quantum chemistry tools accessible to non-experts, thereby accelerating research in chemistry, materials science, and related disciplines. The emphasis on composable actions and integration of literature-based reasoning represents a step toward more general and robust agentic systems for scientific research, and the outlined roadmap could guide future developments in autonomous computational chemistry.

major comments (2)

[Abstract] Abstract: The central claim that Quntur 'supports the full range of calculations available in ORCA 6.0' and 'reasons over software documentation and scientific literature to plan, execute, adapt, and analyze in silico chemistry experiments following best practices' is presented without any quantitative benchmarks, success/failure rates, error analyses, or validation experiments on held-out calculation suites. This absence directly undermines assessment of the core assumption that LLM-driven reasoning will consistently produce correct, best-practice decisions.
[System Design] System architecture and guided research loop description: No concrete case studies, ablation studies on literature retrieval, or catalog of observed failure modes (e.g., invalid input syntax or inappropriate method selection) are provided to demonstrate that the hierarchical agent and composable actions reliably generalize across the ORCA 6.0 feature set without frequent human intervention.

minor comments (1)

[Abstract] Abstract: The phrase 'Despite of its power' is grammatically incorrect and should be changed to 'Despite its power'.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback and recognition of the potential significance of El Agente Quntur. We address each major comment below and will incorporate revisions to strengthen the empirical validation of the system.

read point-by-point responses

Referee: [Abstract] Abstract: The central claim that Quntur 'supports the full range of calculations available in ORCA 6.0' and 'reasons over software documentation and scientific literature to plan, execute, adapt, and analyze in silico chemistry experiments following best practices' is presented without any quantitative benchmarks, success/failure rates, error analyses, or validation experiments on held-out calculation suites. This absence directly undermines assessment of the core assumption that LLM-driven reasoning will consistently produce correct, best-practice decisions.

Authors: We agree that the absence of quantitative benchmarks limits the ability to assess reliability. The manuscript emphasizes the design principles (reasoning-driven decisions, composable actions, and guided research) over exhaustive benchmarking, as the primary contribution is the architectural framework intended to be generalizable. In the revised manuscript, we will add a dedicated validation section including success rates across a held-out suite of ORCA calculations (e.g., geometry optimizations, frequency calculations, and single-point energies), error analyses for common failure modes, and comparisons against expert manual workflows. revision: yes
Referee: [System Design] System architecture and guided research loop description: No concrete case studies, ablation studies on literature retrieval, or catalog of observed failure modes (e.g., invalid input syntax or inappropriate method selection) are provided to demonstrate that the hierarchical agent and composable actions reliably generalize across the ORCA 6.0 feature set without frequent human intervention.

Authors: We acknowledge the need for concrete demonstrations. The current text focuses on the hierarchical structure and research loop to highlight the avoidance of hard-coded policies. In revision, we will include multiple case studies illustrating end-to-end workflows, ablation experiments isolating the contribution of literature retrieval versus documentation-only reasoning, and a catalog of observed failure modes with mitigation strategies (e.g., syntax validation via composable actions and fallback reasoning). These additions will directly address generalization across ORCA features. revision: yes

Circularity Check

0 steps flagged

No circularity: paper describes new software system with no derivations or fitted predictions

full rationale

The manuscript presents the architecture and design principles of El Agente Quntur, a hierarchical multi-agent system for quantum chemistry. It contains no equations, no parameter fitting, no predictions derived from data, and no load-bearing self-citations that reduce the central claims to prior unverified results. The strongest claim is a capability description of a newly constructed tool rather than a derived quantity, so no step reduces to its inputs by construction. This is the expected non-finding for a software-description paper.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The central claim rests on the assumption that current large language models can perform reliable domain-specific reasoning over quantum chemistry literature and ORCA internals without additional training data or fine-tuning details being disclosed.

axioms (1)

domain assumption Large language models can integrate abstract quantum-chemical reasoning across subdisciplines and software syntax
Invoked in the description of guided deep research strategy.

invented entities (1)

El Agente Quntur no independent evidence
purpose: Hierarchical multi-agent research collaborator for quantum chemistry
New system introduced in the paper; no independent prior existence.

pith-pipeline@v0.9.0 · 5636 in / 1236 out tokens · 29581 ms · 2026-05-16T06:45:59.132496+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

hierarchical, multi-agent AI system... elimination of hard-coded procedural policies... guided deep research... supports the full range of calculations available in ORCA 6.0

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

QUASAR: A Universal Autonomous System for Atomistic Simulation and a Benchmark of Its Capabilities
cond-mat.mtrl-sci 2026-01 unverdicted novelty 6.0

QUASAR is a new autonomous LLM-based system that orchestrates multi-scale atomistic simulations and benchmarks as a general reasoning tool rather than a narrow automation script.

Reference graph

Works this paper leans on

231 extracted references · 231 canonical work pages · cited by 1 Pith paper · 4 internal anchors

[1]

Mater.15, 1120–1127 (2016)

Gómez-Bombarelli, R.et al.Design of efficient molecular organic light-emitting diodes by a high-throughput virtual screening and experimental approach.Nat. Mater.15, 1120–1127 (2016)

work page 2016
[2]

& Lin, X

Lin, X., Li, X. & Lin, X. A review on applications of computational methods in drug screening and design. Molecules25, 1375 (2020)

work page 2020
[3]

& Karaman, R

Breijyeh, Z. & Karaman, R. Enzyme models—from catalysis to prodrugs.Molecules26, 3248 (2021)

work page 2021
[4]

Cova, T. F. G. G. & Pais, A. A. C. C. Deep learning for deep chemistry: Optimizing the prediction of chemical patterns.Fchem7(2019)

work page 2019
[5]

Hachmann, J.et al.The harvard clean energy project: Large-scale computational screening and design of organic photovoltaics on the world community grid.J. Phys. Chem. Lett.2, 2241–2251 (2011)

work page 2011
[6]

Sci.7, 698–704 (2014)

Hachmann, J.et al.Lead candidates for high-performance organic photovoltaics from high-throughput quantum chemistry – the harvard clean energy project.Energy Environ. Sci.7, 698–704 (2014)

work page 2014
[7]

J.et al.Biomolecular simulations: From dynamics and mechanisms to computational assays of biological activity.WIREs Comput

Huggins, D. J.et al.Biomolecular simulations: From dynamics and mechanisms to computational assays of biological activity.WIREs Comput. Mol. Sci.9, e1393 (2019)

work page 2019
[8]

& Aspuru-Guzik, A

Andrade, X. & Aspuru-Guzik, A. Real-space Density Functional Theory on graphical processing units: Com- putational approach and comparison to gaussian basis set methods.J. Chem. Theory Comput.9, 4360–4373 (2013)

work page 2013
[9]

Olivares-Amaya, R.et al.Accelerating correlated quantum chemistry calculations using graphical processing units and a mixed precision matrix multiplication library.J. Chem. Theory Comput.6, 135–144 (2010)

work page 2010
[10]

P., Wong, H

de Visser, S. P., Wong, H. P. H., Zhang, Y., Yadav, R. & Sastri, C. V. Tutorial review on the set-up and running of quantum mechanical cluster models for enzymatic reaction mechanisms.Chem. Eur. J.30, e202402468 (2024)

work page 2024
[11]

& Aspuru-Guzik, A

Rebentrost, P., Mohseni, M. & Aspuru-Guzik, A. Role of quantum coherence and environmental fluctuations in chromophoric energy transport.J. Phys. Chem. B113, 9942–9947 (2009)

work page 2009
[12]

Niazi, S. K. Quantum mechanics in drug discovery: A comprehensive review of methods, applications, and future directions.Int. J. Mol. Sci.26, 6325 (2025)

work page 2025
[13]

P.et al.Common workflows for computing material properties using different quantum engines.npj Comput

Huber, S. P.et al.Common workflows for computing material properties using different quantum engines.npj Comput. Mater.7, 136 (2021)

work page 2021
[14]

& Bernales, V

Aspuru-Guzik, A. & Bernales, V. The rise of agents: Computational chemistry is ready for (r)evolution. Polyhedron281, 117707 (2025)

work page 2025
[15]

& Huang, B

Wu, T., Sun, M. & Huang, B. A review of automated workflow pipelines for computational chemists.Small Methods9, 2500308 (2025)

work page 2025
[16]

Curtarolo, S.et al.Aflow: An automatic framework for high-throughput materials discovery.Comput. Mater. Sci.58, 218–226 (2012)

work page 2012
[17]

& Kozinsky, B

Pizzi, G., Cepellotti, A., Sabatini, R., Marzari, N. & Kozinsky, B. AiiDA: automated interactive infrastructure and database for computational science.Comput. Mater. Sci.111, 218–230 (2016)

work page 2016
[18]

Hjorth Larsen, A.et al.The atomic simulation environment—a python library for working with atoms.J. Phys. Condens. Matter.29, 273002 (2017)

work page 2017
[19]

V., Sowndarya S

Alegre-Requena, J. V., Sowndarya S. V., S., Pérez-Soto, R., Alturaifi, T. M. & Paton, R. S. Aqme: Automated quantum mechanical environments for researchers and educators.WIREs Comput. Mol. Sci.13, e1663 (2023)

work page 2023
[20]

Hicks, C. B. & Martinez, T. J. Massively scalable workflows for quantum chemistry: BigChem and ChemCloud. J. Chem. Phys.160, 142501 (2024)

work page 2024
[21]

J., Mori-Sánchez, P

Cohen, A. J., Mori-Sánchez, P. & Yang, W. Challenges for Density Functional Theory.Chem. Rev.112, 289–320 (2012)

work page 2012
[22]

& Grimme, S

Bursch, M., Mewes, J.-M., Hansen, A. & Grimme, S. Best-practice DFT protocols for basic molecular computa- tional chemistry.Angew. Chem. Int. Ed.61, e202205735 (2022). 14

work page 2022
[23]

URLhttp://arxiv.org/abs/2508.14111

Wei, J.et al.From AI for science to agentic science: A survey on autonomous scientific discovery. Preprint at https://arxiv.org/abs/2508.14111 (2025)

work page arXiv 2025
[24]

A., MacKnight, R., Kline, B

Boiko, D. A., MacKnight, R., Kline, B. & Gomes, G. Autonomous chemical research with large language models. Nature624, 570–578 (2023)

work page 2023
[25]

Bran, A.et al.Augmenting large language models with chemistry tools.Nat

M. Bran, A.et al.Augmenting large language models with chemistry tools.Nat. Mach. Intell.6, 525–535 (2024)

work page 2024
[26]

Matter8, 101897 (2025)

Darvish, K.et al.Organa: A robotic assistant for automated chemistry experimentation and characterization. Matter8, 101897 (2025)

work page 2025
[27]

& Moosavi, S

Ansari, M. & Moosavi, S. M. Agent-based learning of materials datasets from the scientific literature.Digit. Discov.3, 2607–2617 (2024)

work page 2024
[28]

D.et al.Cactus: Chemistry agent connecting tool usage to science.ACS Omega9, 46563–46573 (2024)

McNaughton, A. D.et al.Cactus: Chemistry agent connecting tool usage to science.ACS Omega9, 46563–46573 (2024)

work page 2024
[29]

commun.15, 10160 (2024)

Ruan, Y.et al.An automatic end-to-end chemical synthesis development platform powered by large language models.Nat. commun.15, 10160 (2024)

work page 2024
[30]

Dai, T.et al.Autonomous mobile robots for exploratory synthetic chemistry.Nature635, 890–897 (2024)

work page 2024
[31]

W.et al.Chemreasoner: Heuristic search over a large language model’s knowledge space using quantum-chemical feedback

Sprueill, H. W.et al.Chemreasoner: Heuristic search over a large language model’s knowledge space using quantum-chemical feedback. Preprint at https://arxiv.org/abs/2402.10980 (2024)

work page arXiv 2024
[32]

Song, T.et al.A multiagent-driven robotic AI chemist enabling autonomous chemical research on demand.J. Am. Chem. Soc.147, 12534–12545 (2025)

work page 2025
[33]

Preprint at https://arxiv.org/abs/2402.12993 (2025)

Chen, K.et al.ChemMiner: A large language model agent system for chemical literature data mining. Preprint at https://arxiv.org/abs/2402.12993 (2025)

work page arXiv 2025
[34]

Preprint at https://arxiv.org/abs/2501.06590 (2025)

Tang, X.et al.ChemAgent: Self-updating library in large language models improves chemical reasoning. Preprint at https://arxiv.org/abs/2501.06590 (2025)

work page arXiv 2025
[35]

Preprint at https://arxiv.org/abs/2410.07076 (2025)

Yang, Z.et al.MOOSE-Chem: Large language models for rediscovering unseen chemistry scientific hypotheses. Preprint at https://arxiv.org/abs/2410.07076 (2025)

work page arXiv 2025
[36]

The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search

Yamada, Y.et al.The AI scientist-v2: Workshop-level automated scientific discovery via agentic tree search. Preprint at https://arxiv.org/abs/2504.08066 (2025)

work page internal anchor Pith review Pith/arXiv arXiv 2025
[37]

Mitchener, L.et al.Kosmos: An AI scientist for autonomous discovery Preprint at https://arxiv.org/abs/2511.02824 (2025)

work page arXiv 2025
[38]

Preprint at https://arxiv.org/abs/2505.17873 (2025)

Liu, W.et al.MOOSE-Chem3: Toward experiment-guided hypothesis ranking via simulated experimental feedback. Preprint at https://arxiv.org/abs/2505.17873 (2025)

work page arXiv 2025
[39]

ChemAmp: Amplified Chemistry Tools via Composable Agents

Li, Z.et al.ChemHAS: Hierarchical agent stacking for enhancing chemistry tools. Preprint at https://arxiv.org/abs/2505.21569 (2025)

work page internal anchor Pith review Pith/arXiv arXiv 2025
[40]

Preprint at https://arxiv.org/abs/2411.07228 (2025)

Yu, B.et al.ChemToolAgent: The impact of tools on language agents for chemistry problem solving. Preprint at https://arxiv.org/abs/2411.07228 (2025)

work page arXiv 2025
[41]

Preprint at https://arxiv.org/abs/2505.22634 (2025)

Li, R.et al.LabUtopia: High-fidelity simulation and hierarchical benchmark for scientific embodied agents. Preprint at https://arxiv.org/abs/2505.22634 (2025)

work page arXiv 2025
[42]

AI agents in chemical research: GVIM – an intelligent research assistant system.Digit

Ma, K. AI agents in chemical research: GVIM – an intelligent research assistant system.Digit. Discov.4, 355–375 (2025)

work page 2025
[43]

& Ahn, S

Kim, H., Jang, Y. & Ahn, S. MT-Mol: Multi agent system with tool-based reasoning for molecular optimization. Preprint at https://arxiv.org/abs/2505.20820 (2025)

work page arXiv 2025
[44]

Che, X.et al.Csstep: Step-by-step exploration of the chemical space of drug molecules via multi-agent and multi-stage reinforcement learning.Chemical Engineering Science317, 122048 (2025)

work page 2025
[45]

J., Park, N

Callahan, T. J., Park, N. H. & Capponi, S. Agentic mixture-of-workflows for multi-modal chemical search. Preprint at https://arxiv.org/abs/2502.19629 (2025)

work page arXiv 2025
[46]

Preprint at https://arxiv.org/abs/2311.10776 (2025)

Chen, K.et al.Chemist-X: Large language model-empowered agent for reaction condition recommendation in chemical synthesis. Preprint at https://arxiv.org/abs/2311.10776 (2025). 15

work page arXiv 2025
[47]

J.et al.System of Agentic AI for the Discovery of Metal-Organic Frameworks (2025)

Inizan, T. J.et al.System of agentic AI for the discovery of metal-organic frameworks. Preprint at https://arxiv.org/abs/2504.14110 (2025)

work page arXiv 2025
[48]

& Buehler, M

Ghafarollahi, A. & Buehler, M. J. Automating alloy design and discovery with physics-aware multimodal multiagent AI.Proc. Natl. Acad. Sci. U. S. A.122, e2414074122 (2025)

work page 2025
[49]

& Chen, H

Pu, Y., Lin, T. & Chen, H. Piflow: Principle-aware scientific discovery with multi-agent collaboration. Preprint at https://arxiv.org/abs/2505.15047 (2025)

work page arXiv 2025
[50]

& Buehler, M

Ghafarollahi, A. & Buehler, M. J. Autonomous inorganic materials discovery via multi-agent physics-aware scientific reasoning. Preprint at https://arxiv.org/abs/2508.02956 (2025)

work page arXiv 2025
[51]

Preprint at https://arxiv.org/abs/2509.14542 (2025)

Wang, X.et al.S1-MatAgent: A planner driven multi-agent system for material discovery. Preprint at https://arxiv.org/abs/2509.14542 (2025)

work page arXiv 2025
[52]

& Fung, V

Jia, S., Zhang, C. & Fung, V. Llmatdesign: Autonomous materials discovery with large language models. Preprint at https://arxiv.org/abs/2406.13163 (2024)

work page arXiv 2024
[53]

Ding, K.et al.SciToolAgent: a knowledge-graph-driven scientific agent for multitool integration.Nature Computational Science5, 962–972 (2025)

work page 2025
[54]

& Riebesell, J

Chiang, Y., Hsieh, E., Chou, C.-H. & Riebesell, J. LLaMP: Large language model made powerful for high-fidelity materials knowledge retrieval and distillation. Preprint at https://arxiv.org/abs/2401.17244 (2024)

work page arXiv 2024
[55]

Preprint at https://arxiv.org/abs/2507.04053 (2025)

Zhang, B.et al.TopoMAS: Large language model driven topological materials multiagent system. Preprint at https://arxiv.org/abs/2507.04053 (2025)

work page arXiv 2025
[56]

& Calero, S

Petković, M., Menkovski, V. & Calero, S. Towards fully automated molecular simulations: Multi-agent framework for simulation setup and force field extraction. Preprint at https://arxiv.org/abs/2509.10210 (2025)

work page arXiv 2025
[57]

Preprint at https://arxiv.org/abs/2506.05616 (2025)

Zhou, L.et al.Toward greater autonomy in materials discovery agents: Unifying planning, physics, and scientists. Preprint at https://arxiv.org/abs/2506.05616 (2025)

work page arXiv 2025
[58]

Preprint at https://arxiv.org/abs/2512.19458 (2025)

Xia, Z.et al.An agentic framework for autonomous materials computation. Preprint at https://arxiv.org/abs/2512.19458 (2025)

work page arXiv 2025
[59]

Liu, J.et al.VASPilot: MCP-facilitated multi-agent intelligence for autonomous VASP simulations.Chinese Physics B34, 117106 (2025)

work page 2025
[60]

URLhttp://arxiv.org/abs/2507.14267

Wang, Z.et al.DREAMS: Density Functional Theory based research engine for agentic materials simulation. Preprint at https://arxiv.org/abs/2507.14267 (2025)

work page arXiv 2025
[61]

D., Tanikanti, A

Pham, T. D., Tanikanti, A. & Keçeli, M. ChemGraph: An agentic framework for computational chemistry workflows. Preprint at https://arxiv.org/abs/2506.06363 (2025)

work page arXiv 2025
[62]

& Kurban, H

Polat, C., Tuncel, M., Kurban, M., Serpedin, E. & Kurban, H. xChemAgents: Agentic AI for explainable quantum chemistry. Preprint at https://arxiv.org/abs/2505.20574 (2025)

work page arXiv 2025
[63]

Gadde, R. S. K.et al.Chatbot-assisted quantum chemistry for explicitly solvated molecules.Chem. Sci.16, 3852–3864 (2025)

work page 2025
[64]

& Farimani, A

Chandrasekhar, A. & Farimani, A. B. Automating MD simulations for proteins using large language models: NAMD-agent. Preprint at https://arxiv.org/abs/2507.07887 (2025)

work page arXiv 2025
[65]

& White, A

Campbell, Q., Cox, S., Medina, J., Watterson, B. & White, A. D. MDCrow: Automating molecular dynamics workflows with large language models. Preprint at https://arxiv.org/abs/2502.09565 (2025)

work page arXiv 2025
[66]

Rep.15, 10295 (2025)

Shi, Z.et al.A fine-tuned large language model based molecular dynamics agent for code generation to obtain material thermodynamic parameters.Sci. Rep.15, 10295 (2025)

work page 2025
[67]

& Albrecht, S

Masters, C., Grześkiewicz, M. & Albrecht, S. V. Arcane: A multi-agent framework for interpretable and configurable alignment. Preprint at https://arxiv.org/abs/2512.06196 (2025)

work page arXiv 2025
[68]

Zou, Y.et al.El Agente: An autonomous agent for quantum chemistry.Matter8, 102263 (2025)

work page 2025
[69]

& Schwaller, P

Guilbert, S., Masschelein, C., Goumaz, J., Naida, B. & Schwaller, P. Dynamate: An autonomous agent for protein-ligand molecular dynamics simulations. Preprint at https://arxiv.org/abs/2512.10034 (2025)

work page arXiv 2025
[70]

Preprint at https://arxiv.org/abs/2512.06404 (2025)

Soleymanibrojeni, M.et al.Genius: An agentic AI framework for autonomous design and execution of simulation protocols. Preprint at https://arxiv.org/abs/2512.06404 (2025). 16

work page arXiv 2025
[71]

& Tuckerman, M

Lahouari, A., Rogal, J. & Tuckerman, M. E. Automated machine learning pipeline: Large language models- assisted automated data set generation for training machine-learned interatomic potentials.J. Chem. Theory Comput.22, 305–317 (2026)

work page 2026
[72]

& Gong, C

Li, W., Ren, J., Cheng, L. & Gong, C. Autonomous quantum simulation through large language model agents. Preprint at https://arxiv.org/abs/2601.10194 (2026)

work page arXiv 2026
[73]

& Sankaranarayanan, S

Vriza, A., Kornu, U., Koneru, A., Chan, H. & Sankaranarayanan, S. K. R. S. Multi-agentic AI framework for end-to-end atomistic simulations.Digit. Discov.5, 440–452 (2026)

work page 2026
[74]

Preprint at https://arxiv.org/abs/2601.15743 (2026)

Kim, S.et al.Materealize: a multi-agent deliberation system for end-to-end material design and synthesis. Preprint at https://arxiv.org/abs/2601.15743 (2026)

work page arXiv 2026
[75]

QUASAR: A Universal Autonomous System for Atomistic Simulation and a Benchmark of Its Capabilities

Yang, F. & Evans, J. D. QUASAR: A universal autonomous system for atomistic simulation and a benchmark of its capabilities. Preprint at https://arxiv.org/abs/2602.00185 (2026)

work page internal anchor Pith review Pith/arXiv arXiv 2026
[76]

G.et al.El Agente Sólido: A new agent for solid state simulations (2026)

Hari Kumar, S. G.et al.El Agente Sólido: A new agent for solid state simulations (2026). Manuscript in preparation

work page 2026
[77]

Manuscript in preparation

Müller, M.et al.El Agente Fármaco: Agentic AI for autonomous in-silico drug discovery (2026). Manuscript in preparation

work page 2026
[78]

Manuscript in preparation

Choi, C.et al.El Agente Estructural: An agent for structural manipulation of molecular systems (2026). Manuscript in preparation

work page 2026
[79]

Preprint at https://arxiv.org/abs/2512.18847 (2026)

Gustin, I.et al.El Agente Cuántico: Automating quantum simulations. Preprint at https://arxiv.org/abs/2512.18847 (2026)

work page arXiv 2026
[80]

Manuscript in preparation

Bai, J.et al.El Agente Gráfico: Structured Execution Graph for Scientific Agents (2026). Manuscript in preparation

work page 2026

Showing first 80 references.