MoleCode unlocks structural intelligence in large language models
2026/05/15 by Zhiyuan Yan, Chen Liu, Boxuan Zhao +8 · 1 voice
Biochemistry, Genetics and Molecular Biology · Computer Science · #cs.AI #q-bio.BM
paper · pdf · doi:10.48550/arxiv.2605.16480
Abstract
Molecules are graphs, but large language models~(LLMs) are usually asked to reason about them through linear strings. The most popular molecular representation, SMILES, compresses atoms, bonds, branches and rings into a compact sequence in which topology is implicit, forcing LLMs to reconstruct molecular structure before performing the requested chemical operation. Here we introduce MoleCode, an LLM-native, training-free, graph-explicit molecular language in which all molecular components are represented as typed entities with persistent identifiers and explicit relations. MoleCode makes molecular topology directly readable, editable and auditable within the language context, allowing an LLM to operate on structure rather than recover it from syntax. Across molecular reasoning, editing, generation and analysis tasks, this representational shift improves frontier LLMs most strongly when structural access is limiting: unfamiliar molecules, topology-sensitive operations, larger structures and repetitive polymers. It also changes how inference is allocated, replacing long reasoning traces devoted to implicit structural reconstruction with shorter, more chemically directed reasoning over explicit atoms and bonds. In molecular optimization, this enables localized, property-aligned edits that preserve structural similarity to the starting compounds. The same Subgraph--Node--Edge grammar extends beyond small molecules to polymers, Markush structures, mechanism-style transformations and interleaved scientific documents, including research articles and patent disclosures in which chemical information is distributed across text and images. These results suggest that the interface between scientific objects and LLMs should not treat structure as something to be decoded from text. When the object of reasoning is relational, the structure itself should be part of the language.
Citations
- OpenAI GPT-5 System Card
- FGBench: A Dataset and Benchmark for Molecular Property Reasoning at Functional Group-Level in Large Language Models
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Beyond Chemical QA: Evaluating LLM's Chemical Reasoning with Modular Chemical Operations
- Improving Chemical Understanding of LLMs via SMILES Parsing
- MolLangBench: A Comprehensive Benchmark for Language-Prompted Molecular Structure Recognition, Editing, and Generation
- Assessing the Chemical Intelligence of Large Language Models
- How to Detect and Defeat Molecular Mirage: A Metric-Driven Benchmark for Hallucination in LLM-based Molecular Comprehension
- Mol-LLaMA: Towards General Understanding of Molecules in Large Molecular Language Model
- DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning
- Navigating Chemical-Linguistic Sharing Space with Heterogeneous Molecular Encoding
- Speak-to-Structure: Evaluating LLMs in Open-domain Natural Language-Driven Molecule Generation
- MolParser: End-to-end Visual Recognition of Molecule Structures in the Wild
- LLaMo: Large Language Model-based Molecular Graph Assistant
- UniCorn: A Unified Contrastive Learning Approach for Multi-view Molecular Representation Learning
- LlaSMol: Advancing Large Language Models for Chemistry with a Large-Scale, Comprehensive, High-Quality Instruction Tuning Dataset
- Towards 3D Molecule-Text Interpretation in Language Models
- InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery
- MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adapter
- BioT5: Enriching Cross-modal Integration in Biology with Chemical Knowledge and Natural Language Associations
- Multimodal Molecular Pretraining via Modality Blending
- Mol-Instructions: A Large-Scale Biomolecular Instruction Dataset for Large Language Models
- MolXPT: Wrapping Molecules with Text for Generative Pre-training
- Enhancing Activity Prediction Models in Drug Discovery with the Ability to Understand Human Language
- t-SMILES: A Scalable Fragment-based Molecular Representation Framework for De Novo Molecule Generation
- Multi-modal Molecule Structure-text Model for Text-based Retrieval and Editing
- PubChem 2023 update
- A Molecular Multimodal Foundation Model Associating Molecule Graphs with Natural Language
- Graph neural networks for materials science and chemistry
- Translation between Molecules and Natural Language
- BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
- Pre-training Molecular Graph Representation with 3D Geometry
- GraphCL: Contrastive Self-Supervised Learning of Graph Representations
- Junction Tree Variational Autoencoder for Molecular Graph Generation
- SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
- Neural Message Passing for Quantum Chemistry
- Grammar Variational Autoencoder
- Semi-Supervised Classification with Graph Convolutional Networks
- The Nature of External Representations in Problem Solving
- SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules
- What can Large Language Models do in chemistry? A comprehensive benchmark on eight tasks
- Representations in Distributed Cognitive Tasks
Discussions
- Yan, Liu, Zhao, Lin, Zhao, Wang, Lv, Li, Zhang, Yuan, Mo: MoleCode unlocks structural intelligence in large language models https://arxiv.org/abs/2605.16480 https://arxiv.org/pdf/2605.16480 https://ar [bsky, 0 points, 0 comments]
Related