Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints

Hugging Face Daily Papers Papers
benchmarking llms molecular-generation drug-design spatial-constraints 3d-molecules structure-based-drug-design

Summary

This paper benchmarks general-purpose LLMs against specialized diffusion models for generating binding molecules under 3D spatial constraints, finding that LLMs show promise despite currently lagging behind state-of-the-art approaches.

Structure-based drug design (SBDD) leverages the 3D structure of protein targets, often complemented by other spatial constraints, to generate candidate binding molecules. While diffusion models have dominated as a leading paradigm for high-quality 3D molecule generation, LLM-based methods are rapidly emerging in molecular design and have shown competitive performance in pocket-conditioned molecular generation. However, their ability to reason about physics and 3D spatial environments is largely underexplored. In this work, we systematically analyze whether current general-purpose LLMs are capable of navigating complex 3D constraints compared to established baselines such as specialized diffusion models. We consider 3D ligand generation conditioned on protein pockets together with ligand- and interaction-derived spatial constraints, including anchor fragments, pharmacophore points, and mandatory pocket-ligand interactions. To enable this evaluation, we introduce 3D-Fit - a token-efficient benchmarking strategy for assessing LLM performance on multi-conditioned spatial molecule generation. Our findings reveal a clear pattern in LLM spatial capabilities: while they still lag behind state-of-the-art approaches, they are promising and can handle multiple spatial constraints simultaneously, enabling scaling to heterogeneous setups.
Original Article
View Cached Full Text

Cached at: 07/21/26, 02:37 PM

Paper page - Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints

Source: https://huggingface.co/papers/2607.18144

Abstract

Structure-baseddrugdesign(SBDD)leveragesthe3Dstructureofproteintargets,oftencomplementedbyotherspatialconstraints,togeneratecandidatebindingmolecules.Whilediffusionmodelshavedominatedasaleadingparadigmforhigh-quality3Dmoleculegeneration,LLM-basedmethodsarerapidlyemerginginmoleculardesignandhaveshowncompetitiveperformanceinpocket-conditionedmoleculargeneration.However,theirabilitytoreasonaboutphysicsand3Dspatialenvironmentsislargelyunderexplored.Inthiswork,wesystematicallyanalyzewhethercurrentgeneral-purposeLLMsarecapableofnavigatingcomplex3Dconstraintscomparedtoestablishedbaselinessuchasspecializeddiffusionmodels.Weconsider3Dligandgenerationconditionedonproteinpocketstogetherwithligand-andinteraction-derivedspatialconstraints,includinganchorfragments,pharmacophorepoints,andmandatorypocket-ligandinteractions.Toenablethisevaluation,weintroduce3D-Fit-atoken-efficientbenchmarkingstrategyforassessingLLMperformanceonmulti-conditionedspatialmoleculegeneration.OurfindingsrevealaclearpatterninLLMspatialcapabilities:whiletheystilllagbehindstate-of-the-artapproaches,theyarepromisingandcanhandlemultiplespatialconstraintssimultaneously,enablingscalingtoheterogeneoussetups.

View arXiv pageView PDFGitHub0Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.18144 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.18144 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.18144 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

MolDesignBench: Evaluating LLM-based Agent for Scenario-grounded Molecular Design

arXiv cs.AI

MolDesignBench is a new benchmark for evaluating LLM-based agents in scenario-grounded molecular design, comprising 2K instances with implicit and explicit constraints, revealing that current frontier LLMs achieve low success rates, especially in reasoning about implicit constraints and infeasibility detection.

Diffusion Language Models: An Experimental Analysis

arXiv cs.AI

A systematic experimental analysis evaluating eight state-of-the-art Diffusion Language Models across multiple benchmarks, analyzing trade-offs between generation quality and computational efficiency.

Assessing Language Models for Scientific Applications: Inertial Confinement Fusion as a Use Case

arXiv cs.AI

Researchers from Los Alamos National Laboratory and NVIDIA present an expert-curated benchmark of 250 questions on inertial confinement fusion to evaluate LLMs' scientific fluency, physical reasoning, and quantitative analysis. They compare frontier open/closed models with Prospero, a domain-specialized model, finding that specialization improves scores but critical physics errors persist, and recommend assessing existing capabilities and retrieval benefits before investing in further specialization.