A framework for evaluating the chemical knowledge and reasoning abilities of large language models against the expertise of chemists Adrian Mirza, Nawaf Alampara, Sreekanth Kunchapu, Martiño Ríos-García, Benedict Emoekabu, et al. Nature Chemistry, 2025 Large language models (LLMs) have gained widespread interest owing to their ability to process human language and perform tasks on which they have not been explicitly trained. However, we possess only a limited systematic understanding of the chemical capabilities of LLMs, which would be required to improve models and mitigate potential harm. Here we introduce ChemBench, an automated framework for evaluating the chemical knowledge and reasoning abilities of state-of-the-art LLMs against the expertise of chemists. We curated more than 2,700 question–answer pairs, evaluated leading open- and closed-source LLMs and found that the best models, on average, outperformed the best human chemists in our study. However, the models struggle with some basic tasks and provide overconfident predictions. These findings reveal LLMs’ impressive chemical capabilities while emphasizing the need for further research to improve their safety and usefulness. They also suggest adapting chemistry education and show the value of benchmarking frameworks for evaluating LLMs in specific domains.
MOFGalaxyNet: a social network analysis for predicting guest accessibility in metal–organic frameworks utilizing graph convolutional networks Mehrdad Jalali, A. D. Dinga Wonanke, Christof Wöll Journal of Cheminformatics, 2023 Metal–organic frameworks (MOFs), are porous crystalline structures comprising of metal ions or clusters intricately linked with organic entities, displaying topological diversity and effortless chemical flexibility. These characteristics render them apt for multifarious applications such as adsorption, separation, sensing, and catalysis. Predominantly, the distinctive properties and prospective utility of MOFs are discerned post-manufacture or extrapolation from theoretically conceived models. For empirical researchers unfamiliar with hypothetical structure development, the meticulous crystal engineering of a high-performance MOF for a targeted application via a bottom-up approach resembles a gamble. For example, the precise pore limiting diameter (PLD), which determines the guest accessibility of any MOF cannot be easily inferred with mere knowledge of the metal ion and organic ligand. This limitation in bottom-up conceptual understanding of specific properties of the resultant MOF may contribute to the cautious industrial-scale adoption of MOFs.Consequently, in this study, we take a step towards circumventing this limitation by designing a new tool that predicts the guest accessibility—a MOF key performance indicator—of any given MOF from information on only the organic linkers and the metal ions. This new tool relies on clustering different MOFs in a galaxy-like social network, MOFGalaxyNet, combined with a Graphical Convolutional Network (GCN) to predict the guest accessibility of any new entry in the social network. The proposed network and GCN results provide a robust approach for screening MOFs for various host–guest interaction studies.
Role of Host-Guest Interaction in Understanding Polymerisation in Metal-Organic Frameworks A.D. Dinga Wonanke, Poppy Bennett, Lewis Caldwell, Matthew A. Addicoat Frontiers in Chemistry, 2021 Metal-organic frameworks, MOFs, offer an effective template for polymerisation of polymers with precisely controlled structures within the sub-nanometre scales. However, synthetic difficulties such as monomer infiltration, detailed understanding of polymerisation mechanisms within the MOF nanochannels and the mechanism for removing the MOF template post polymerisation have prevented wide scale implementation of polymerisation in MOFs. This is partly due to the significant lack in understanding of the energetic and atomic-scale intermolecular interactions between the monomers and the MOFs. Consequently in this study, we explore the interaction of varied concentration of styrene, and 3,4-ethylenedioxythiophene (EDOT), at the surface and in the nanochannel of Zn 2 (1,4-ndc) 2 (dabco), where 1,4-ndc = 1,4-naphthalenedicarboxylate and dabco = 1,4-diazabicyclo[2.2.2]octane. Our results showed that the interactions between monomers are stronger in the nanochannels than at the surfaces of the MOF. Moreover, the MOF-monomer interactions are strongest in the nanochannels and increase with the number of monomers. However, as the number of monomers increases, the monomers turn to bind more strongly at the surface leading to a potential agglomeration of the monomers at the surface.
FAIR-MOFs: Structure-centred synthesis inference from three-dimensional structures of metal-organic frameworks T Heine, D Wonanke, A Longa, A Pankajakshan, O Joseph, L Himanen, ... 2026
FAIR-MOFs: A Comprehensive Database for Accelerating the Discovery and Synthesis of Metal-Organic Frameworks D Wonanke, A Longa, A Pankajakshan, L Himanen, AN Ladines, ... ChemRxiv 2025 (1024) , 2025 2025 Citations: 1
The Black Hole Strategy: Gravity-Based Representative Sampling for Frugal Graph Learning on Metal–Organic Framework Networks M Jalali, ADD Wonanke, P Friederich, C Wöll Journal of Chemical Information and Modeling 65 (20), 10885-10902 , 2025 2025 Citations: 4
A framework for evaluating the chemical knowledge and reasoning abilities of large language models against the expertise of chemists A Mirza, N Alampara, S Kunchapu, M Ríos-García, B Emoekabu, ... Nature Chemistry 17 (7), 1027-1034 , 2025 2025 Citations: 115
Engineering photoswitching dynamics in 3D photochromic metal–organic frameworks through a metal–organic polyhedron design E Jin, V Bon, S Das, ADD Wonanke, M Etter, MA Karlsen, A De, ... Journal of the American Chemical Society 147 (10), 8568-8577 , 2025 2025 Citations: 19
Dataset for A framework for evaluating the chemical knowledge and reasoning abilities of large language models against the expertise of chemists A Mirza, N Alampara, S Kunchapu, M Ríos-García, B Emoekabu, ... 2025
Are large language models superhuman chemists? A Mirza, N Alampara, S Kunchapu, M Ríos-García, B Emoekabu, ... arXiv preprint arXiv:2404.01475 , 2024 2024 Citations: 99
Correction: MOFGalaxyNet: a social network analysis for predicting guest accessibility in metal–organic frameworks utilizing graph convolutional networks M Jalali, ADD Wonanke, C Wöll Journal of Cheminformatics 15, 108 , 2023 2023 Citations: 2
MOFGalaxyNet: a social network analysis for predicting guest accessibility in metal–organic frameworks utilizing graph convolutional networks M Jalali, ADD Wonanke, C Wöll Journal of Cheminformatics 15 (1), 94 , 2023 2023 Citations: 21
Covalent organic framework as a metal-free photocatalyst for dye degradation and radioactive iodine adsorption S Ruidas, A Chowdhury, A Ghosh, A Ghosh, S Mondal, ADD Wonanke, ... Langmuir 39 (11), 4071-4081 , 2023 2023 Citations: 69
A Social Network-Guided Approach to Machine Learning for Metal-Organic Framework Property Prediction M Jalali, AD Wonanke, C Woll 5th European Conference on Metal Organic Frameworks and Porous Polymers … , 2023 2023
Ionic covalent organic nanosheet (iCON)–quaternized polybenzimidazole nanocomposite anion-exchange membranes to enhance the performance of membrane capacitive deionization R McNair, S Kumar, ADD Wonanke, MA Addicoat, RAW Dryfe, G Szekely Desalination 533, 115777 , 2022 2022 Citations: 28
Hydroxyl-functionalized covalent organic framework membranes: fast organic solvent nanofiltration DB Shinde, L Cao, S Kumar, Z Zhou, I Chen, D Wonanke, M Addicoat, ... Journal of Membrane Science and Research 8 (3) , 2022 2022 Citations: 3
Prediction of anharmonic, condensed-phase IR spectra using a composite approach: Discrete encapsulated chloride hydrates ADD Wonanke, DL Crittenden Journal of Molecular Spectroscopy 387, 111660 , 2022 2022
Norbornane-based covalent organic frameworks for gas separation S Kumar, MA Abdulhamid, ADD Wonanke, MA Addicoat, G Szekely Nanoscale 14 (6), 2475-2481 , 2022 2022 Citations: 46
Effect of unwanted guest molecules on the stacking configuration of covalent organic frameworks: a periodic energy decomposition analysis ADD Wonanke, MA Addicoat Physical Chemistry Chemical Physics 24 (25), 15494-15501 , 2022 2022 Citations: 10
Tailored pore size and microporosity of covalent organic framework (COF) membranes for improved molecular separation DB Shinde, L Cao, X Liu, DAD Wonanke, Z Zhou, MN Hedhili, M Addicoat, ... Journal of Membrane Science Letters 1 (2), 100008 , 2021 2021 Citations: 30
A dual-function highly crystalline covalent organic framework for HCl sensing and visible-light heterogeneous photocatalysis Y Nailwal, ADD Wonanke, MA Addicoat, SK Pal Macromolecules 54 (13), 6595-6604 , 2021 2021 Citations: 46
Supramolecular Chromatographic Separation of C 60 and C 70 Fullerenes: Flash Column Chromatography vs. High Pressure Liquid Chromatography S Mekapothula, ADD Wonanke, MA Addicoat, DJ Boocock, JD Wallis, ... International Journal of Molecular Sciences 22 (11), 5726 , 2021 2021 Citations: 5
Role of Host-Guest Interaction in Understanding Polymerisation in Metal-Organic Frameworks W A.D. Dinga, B Poppy, C Lewis, A Matthew, A. Frontiers in Chemistry 9, 716294 , 2021 2021 Citations: 3
MOST CITED SCHOLAR PUBLICATIONS
Confining H 3 PO 4 network in covalent organic frameworks enables proton super flow S Tao, L Zhai, AD Dinga Wonanke, MA Addicoat, Q Jiang, D Jiang Nature Communications 11 (1), 1981 , 2020 2020.0 Citations: 190
Pore engineering of ultrathin covalent organic framework membranes for organic solvent nanofiltration and molecular sieving DB Shinde, L Cao, ADD Wonanke, X Li, S Kumar, X Liu, MN Hedhili, ... Chemical science 11 (21), 5434-5440 , 2020 2020.0 Citations: 124
A framework for evaluating the chemical knowledge and reasoning abilities of large language models against the expertise of chemists A Mirza, N Alampara, S Kunchapu, M Ríos-García, B Emoekabu, ... Nature Chemistry 17 (7), 1027-1034 , 2025 2025.0 Citations: 115
Are large language models superhuman chemists? A Mirza, N Alampara, S Kunchapu, M Ríos-García, B Emoekabu, ... arXiv preprint arXiv:2404.01475 , 2024 2024.0 Citations: 99
Covalent organic framework as a metal-free photocatalyst for dye degradation and radioactive iodine adsorption S Ruidas, A Chowdhury, A Ghosh, A Ghosh, S Mondal, ADD Wonanke, ... Langmuir 39 (11), 4071-4081 , 2023 2023.0 Citations: 69
Norbornane-based covalent organic frameworks for gas separation S Kumar, MA Abdulhamid, ADD Wonanke, MA Addicoat, G Szekely Nanoscale 14 (6), 2475-2481 , 2022 2022.0 Citations: 46
A dual-function highly crystalline covalent organic framework for HCl sensing and visible-light heterogeneous photocatalysis Y Nailwal, ADD Wonanke, MA Addicoat, SK Pal Macromolecules 54 (13), 6595-6604 , 2021 2021.0 Citations: 46
Tailored pore size and microporosity of covalent organic framework (COF) membranes for improved molecular separation DB Shinde, L Cao, X Liu, DAD Wonanke, Z Zhou, MN Hedhili, M Addicoat, ... Journal of Membrane Science Letters 1 (2), 100008 , 2021 2021.0 Citations: 30
Ionic covalent organic nanosheet (iCON)–quaternized polybenzimidazole nanocomposite anion-exchange membranes to enhance the performance of membrane capacitive deionization R McNair, S Kumar, ADD Wonanke, MA Addicoat, RAW Dryfe, G Szekely Desalination 533, 115777 , 2022 2022.0 Citations: 28
MOFGalaxyNet: a social network analysis for predicting guest accessibility in metal–organic frameworks utilizing graph convolutional networks M Jalali, ADD Wonanke, C Wöll Journal of Cheminformatics 15 (1), 94 , 2023 2023.0 Citations: 21
Engineering photoswitching dynamics in 3D photochromic metal–organic frameworks through a metal–organic polyhedron design E Jin, V Bon, S Das, ADD Wonanke, M Etter, MA Karlsen, A De, ... Journal of the American Chemical Society 147 (10), 8568-8577 , 2025 2025.0 Citations: 19
Effect of unwanted guest molecules on the stacking configuration of covalent organic frameworks: a periodic energy decomposition analysis ADD Wonanke, MA Addicoat Physical Chemistry Chemical Physics 24 (25), 15494-15501 , 2022 2022.0 Citations: 10
Are large language models superhuman chemists?, 2024 A Mirza, N Alampara, S Kunchapu, M Rıos-Garcıa, B Emoekabu, ... URL https://arxiv. org/abs/2404.01475 , 0 Citations: 7
Are large language models superhuman chemists?, arXiv, 2024 A Mirza, N Alampara, S Kunchapu, B Emoekabu, A Krishnan, M Wilhelmi, ... arXiv preprint arXiv:2404.01475 10 , 0 Citations: 7
Predicting the Outcome of Photocyclisation Reactions: A Joint Experimental and Computational Investigation ADD Wonanke, JL Ferguson, CM Fitchett, DL Crittenden Chemistry–An Asian Journal 14 (8), 1293-1303 , 2019 2019.0 Citations: 6
Supramolecular Chromatographic Separation of C 60 and C 70 Fullerenes: Flash Column Chromatography vs. High Pressure Liquid Chromatography S Mekapothula, ADD Wonanke, MA Addicoat, DJ Boocock, JD Wallis, ... International Journal of Molecular Sciences 22 (11), 5726 , 2021 2021.0 Citations: 5
Beyond the Woodward-Hoffman rules: what controls reactivity in eliminative aromatic ring-forming reactions? ADD Wonanke, DL Crittenden Australian Journal of Chemistry 71 (4), 249-256 , 2018 2018.0 Citations: 5
The Black Hole Strategy: Gravity-Based Representative Sampling for Frugal Graph Learning on Metal–Organic Framework Networks M Jalali, ADD Wonanke, P Friederich, C Wöll Journal of Chemical Information and Modeling 65 (20), 10885-10902 , 2025 2025.0 Citations: 4
Hydroxyl-functionalized covalent organic framework membranes: fast organic solvent nanofiltration DB Shinde, L Cao, S Kumar, Z Zhou, I Chen, D Wonanke, M Addicoat, ... Journal of Membrane Science and Research 8 (3) , 2022 2022.0 Citations: 3
Role of Host-Guest Interaction in Understanding Polymerisation in Metal-Organic Frameworks W A.D. Dinga, B Poppy, C Lewis, A Matthew, A. Frontiers in Chemistry 9, 716294 , 2021 2021.0 Citations: 3