<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cao-etal-2026-molecule">
    <titleInfo>
        <title>Molecule Generation through Reasoning with Large Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Congfeng</namePart>
        <namePart type="family">Cao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Di</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhi</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jia-Hong</namePart>
        <namePart type="family">Huang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jelke</namePart>
        <namePart type="family">Bloem</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Khalil</namePart>
        <namePart type="family">Sima’an</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yue</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ruifeng</namePart>
            <namePart type="family">Xu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ayu</namePart>
            <namePart type="family">Purwarianti</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tong</namePart>
            <namePart type="family">Rong</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jun</namePart>
            <namePart type="family">Suzuki</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Hengqin, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-444-6</identifier>
    </relatedItem>
    <abstract>Molecule generation is significant for its potential in scientific discovery and practical applications. Recent attempts often frame this task as a *translation* problem from molecular caption to structural representation, such as SMILES. This paper first examines the feasibility of modeling the task as a reasoning process with large language models (LLMs), generating higher-quality molecules through structural decomposition and recombination within curated Chain-of-Thought (CoT). We then introduce a workflow for curating accurate CoT data, incorporating both machine and expert verification. Lastly, we demonstrate that with a limited dataset of 4,213 curated samples, namely **MolCoT4K**, we elicit strong reasoning capabilities for molecule generation in open-source LLMs such as Qwen2.5-7B, achieving state-of-the-art exact match accuracy over strong open-source baselines (e.g., MolT5 and LlaSMol) as well as advanced commercial LLMs like GPT-4o. Moreover, the resulting model, **MolGeneration**, attains a Pass@16 exact match accuracy of 48.46%, highlighting its strong potential for real-world experimental applications when supported by a feasible external verifier or chemistry experts. Our analysis shows that the correctness of the CoT path is crucial, while reasoning ability primarily enhances accuracy in fine-grained molecule generation.</abstract>
    <identifier type="citekey">cao-etal-2026-molecule</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-aacl.195/</url>
    </location>
    <part>
        <date>2026-11</date>
        <extent unit="page">
            <start>3862</start>
            <end>3879</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Molecule Generation through Reasoning with Large Language Models
%A Cao, Congfeng
%A Wu, Di
%A Zhang, Zhi
%A Huang, Jia-Hong
%A Bloem, Jelke
%A Sima’an, Khalil
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F cao-etal-2026-molecule
%X Molecule generation is significant for its potential in scientific discovery and practical applications. Recent attempts often frame this task as a *translation* problem from molecular caption to structural representation, such as SMILES. This paper first examines the feasibility of modeling the task as a reasoning process with large language models (LLMs), generating higher-quality molecules through structural decomposition and recombination within curated Chain-of-Thought (CoT). We then introduce a workflow for curating accurate CoT data, incorporating both machine and expert verification. Lastly, we demonstrate that with a limited dataset of 4,213 curated samples, namely **MolCoT4K**, we elicit strong reasoning capabilities for molecule generation in open-source LLMs such as Qwen2.5-7B, achieving state-of-the-art exact match accuracy over strong open-source baselines (e.g., MolT5 and LlaSMol) as well as advanced commercial LLMs like GPT-4o. Moreover, the resulting model, **MolGeneration**, attains a Pass@16 exact match accuracy of 48.46%, highlighting its strong potential for real-world experimental applications when supported by a feasible external verifier or chemistry experts. Our analysis shows that the correctness of the CoT path is crucial, while reasoning ability primarily enhances accuracy in fine-grained molecule generation.
%U https://aclanthology.org/2026.findings-aacl.195/
%P 3862-3879

Markdown (Informal)

[Molecule Generation through Reasoning with Large Language Models](https://aclanthology.org/2026.findings-aacl.195/) (Cao et al., Findings 2026)

• Molecule Generation through Reasoning with Large Language Models (Cao et al., Findings 2026)

ACL

• Congfeng Cao, Di Wu, Zhi Zhang, Jia-Hong Huang, Jelke Bloem, and Khalil Sima’an. 2026. Molecule Generation through Reasoning with Large Language Models. In Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026, pages 3862–3879, Hengqin, China. Association for Computational Linguistics.