MolReasoner: Toward Effective and Interpretable Reasoning for Molecular LLMs

2025/08/04 by Guojiang Zhao, Zixiang Lu, Zhao, Guojiang +27 · 3 citations
Computer Science · Materials Science · #Advanced Graph Neural Networks #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Computational Drug Discovery Methods #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning in Materials Science

paper · pdf · doi:10.48550/arxiv.2508.02066

openalex publication_date 2025/08/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Large Language Models (LLMs) have shown impressive performance across various domains, but their ability to perform molecular reasoning remains underexplored. Existing methods mostly rely on general-purpose prompting, which lacks domain-specific molecular semantics, or fine-tuning, which faces challenges in interpretability and reasoning depth, often leading to structural and textual hallucinations. To address these issues, we introduce MolReasoner, a two-stage framework that transitions LLMs from memorization to high-fidelity chemical reasoning. In the Mol-SFT stage, knowledge-enhanced Chain-of-Thought (CoT) data provides a strong foundation, while the Mol-RL stage refines reasoning using a novel, task-adaptive reward system to mitigate hallucinations. Extensive evaluations demonstrate that MolReasoner significantly outperforms a wide range of strong baselines in both molecule generation and captioning tasks. Further analyses highlight the framework's synergistic design and its ability to produce more interpretable outputs. Our work presents a principled and effective new approach for advancing high-fidelity molecular reasoning.

Citations

• Training a Scientific Reasoning Model for Chemistry
• Qwen3 Technical Report
• Fin-R1: A Large Language Model for Financial Reasoning through Reinforcement Learning
• Mol-LLaMA: Towards General Understanding of Molecules in Large Molecular Language Model
• SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training
• Structural Reasoning Improves Molecular Understanding of LLM
• CBGBench: Fill in the Blank of Protein-Molecule Complex Binding Graph
• BioT5+: Towards Generalized Biological Understanding with IUPAC Integration and Multi-task Tuning
• LlaSMol: Advancing Large Language Models for Chemistry with a Large-Scale, Comprehensive, High-Quality Instruction Tuning Dataset
• BioXP-0.5B: Explainable Medical-AI via RL-GRPO
• Towards 3D Molecule-Text Interpretation in Language Models
• MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adapter
• What can Large Language Models do in chemistry? A comprehensive benchmark on eight tasks
• Multi-modal Molecule Structure-text Model for Text-based Retrieval and Editing
• Galactica: A Large Language Model for Science
• A Molecular Multimodal Foundation Model Associating Molecule Graphs with Natural Language
• Training language models to follow instructions with human feedback
• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
• HellaSwag: Can a Machine Really Finish Your Sentence?
• Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering
• Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
• Proximal Policy Optimization Algorithms
• Crowdsourcing Multiple Choice Science Questions
• Computational Methods in Drug Discovery

Cited by

• A Scientific Reasoning Model for Organic Synthesis Procedure Generation
• Coder as Editor: Code-driven Interpretable Molecular Optimization
• DrugReasoner: Interpretable Drug Approval Prediction with a Reasoning-augmented Language Model

Related

• Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery
• Training a Scientific Reasoning Model for Chemistry
• Mol-LLaMA: Towards General Understanding of Molecules in Large Molecular Language Model
• AttriLens-Mol: Attribute Guided Reinforcement Learning for Molecular Property Prediction with Large Language Models
• ChemATP: A Training-Free Chemical Reasoning Framework for Large Language Models