大模型
arxiv.org ·
分子几何理解已在前沿大型语言模型中意外涌现
该 arXiv 预印本测试了现代 LLM 对分子三维结构的理解能力，使用 27 个小有机分子的 810 个构象几何与 DFT 能量。结果显示，2025 年 7 月前发布的模型难以按稳定性排序构象，而近期前沿模型（文中称 GPT-5.6 Sol、Kimi K3、Gemini 3.6 Flash）达到有竞争力的准确率并超过 Universal Force Field，GPT-6 Astra 与 Claude Opus 5 接近 GFN-FF。分析表明模型在氢键等有明确科学语言描述的分子上表现最佳，在环张力等定义模糊的概念上弱于力场；该能力与科学、编程和抽象推理基准表现强相关，可能是在相关但概念不同的任务训练中意外涌现。作者认为这为 AI 驱动的分子设计提供了现实起点。
AI推荐理由 arXiv 预印本，作者来自俄罗斯有机化学研究所、HSE 与佛罗里达州立大学，方法可核查：在 27 个小有机分子的 810 个构象几何与 DFT 能量上测试 LLM 的构象稳定性排序能力。新信息是 2025 年 7 月后发布的前沿模型（文中称 GPT-5.6 Sol、Kimi K3、Gemini 3.6 Flash）排序精度超过 Universal Force Field，GPT-6 Astra 与 Claude Opus 5 接近 GFN-FF；并指出模型在氢键主导分子上表现好、在环张力等模糊概念上弱，且该能力与科学/编程/抽象推理基准表现相关。局限：预印本未经同行评审，模型名称与结论需以正式发表为准，摘要未给出具体误差数值。
78
正文
Physics > Chemical Physics
arXiv:2609.20666 (physics)
[Submitted on 17 Sep 2026]
Title:Molecular Geometry Understanding Has Unintendedly Emerged in Frontier Large Language Models
Authors:Gregorii A. Semakin (1 and 2), Timofey V. Losev (1), Ilya V. Prolomov (1), Stepan N. Ostarkov (1), Igor V. Alabugin (3), Michael G. Medvedev (1) ((1) Zelinsky Institute of Organic Chemistry RAS, (2) HSE University, (3) Florida State University)
View a PDF of the paper titled Molecular Geometry Understanding Has Unintendedly Emerged in Frontier Large Language Models, by Gregorii A. Semakin (1 and 2) and 7 other authors
View PDF
HTML (experimental)
Abstract:Large language models (LLMs) have already shown strong capabilities in solving complex chemical problems expressed in natural language. Yet, many tasks performed by chemists require understanding of 3D structures of chemical compounds and reasoning about them -- abilities, which, to the best of our knowledge, were neither intended by LLM developers nor tested in contemporary models. Such understanding is essential for autonomous molecular discovery and drug development, which is one of the Holy Grails of AI application to chemistry. We tested this capability in modern LLMs on 810 geometries and DFT energies of 27 small organic molecules. Strikingly, while models released before July 2025 struggled to rank conformers by stability, many recent frontier models, including GPT-5.6 Sol, Kimi K3, and Gemini 3.6 Flash achieved competitive accuracy, outperforming the Universal Force Field, and GPT-6 Astra and Claude Opus 5 came very close to a modern GFN-FF force field. Analysis of models' explanations for their rankings suggests that they perform best for molecules with well-defined intramolecular interactions -- hydrogen bonds -- which are well quantified in scientific language; at the same time all models except the two newest -- GPT-6 Astra and Claude Opus 5 -- struggle with molecules governed by loosely defined concepts (e.g., ring strain), where force fields excel, suggesting that the scientific language itself might impose constraints on LLMs. Notably, a model's ability to rank conformers is strongly associated with its performance on scientific, coding, and abstract-reasoning benchmarks, suggesting that it emerged unintendedly from models training on linked but conceptually different tasks. Thanks to this generalization of mere scientific knowledge into understanding molecular structures, current frontier models offer a realistic starting point for AI-driven molecular design.
Comments:
Subjects:
Chemical Physics (physics.chem-ph)
Cite as:
arXiv:2609.20666 [physics.chem-ph]
(or
arXiv:2609.20666v1 [physics.chem-ph] for this version)
https://doi.org/10.48550/arXiv.2609.20666
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Michael G. Medvedev [view email]
[v1]
Thu, 17 Sep 2026 16:42:41 UTC (5,638 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Molecular Geometry Understanding Has Unintendedly Emerged in Frontier Large Language Models, by Gregorii A. Semakin (1 and 2) and 7 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
physics.chem-ph
new
|
recent
| 2026-09
Change to browse by:
physics
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)