mlsanityData to 14 Oct 2026

NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation

arXiv 2502.12638 ↗First posted 18 Feb 2025q-bio.QM cs.LG q-bio.BMcited by 17

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

3D molecule generation is crucial for drug discovery and material design. While prior efforts focus on 3D diffusion models for their benefits in modeling continuous 3D conformers, they overlook the advantages of 1D SELFIES-based Language Models (LMs), which can generate 100% valid molecules and leverage the billion-scale 1D molecule datasets. To combine these advantages for 3D molecule generation, we propose a foundation model -- NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation. NExT-Mol uses an extensively pretrained molecule LM for 1D molecule generation, and subsequently predicts the generated molecule's 3D conformers with a 3D diffusion model. We enhance NExT-Mol's performance by scaling up the LM's model size, refining the diffusion neural architecture, and applying 1D to 3D transfer learning. Notably, our 1D molecule LM significantly outperforms baselines in distributional similarity while ensuring validity, and our 3D diffusion model achieves leading performances in conformer prediction. Given these improvements in 1D and 3D modeling, NExT-Mol achieves a 26% relative improvement in 3D FCD for de novo 3D generation on GEOM-DRUGS, and a 13% average relative gain for conditional 3D generation on QM9-2014. Our codes and pretrained checkpoints are available at https://github.com/acharkq/NExT-Mol.

Key facts. Cited by 17 papers in the map (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 82 of its 115 references that are in the map. First posted on arXiv 18 Feb 2025.

Builds on

82 of 115 references.

Closest ancestors · 9

Same direction, newest first

  1. 2024Text-guided Diffusion Model for 3D Molecule Generation4 Oct
  2. 2023Swallowing the Bitter Pill: Simplified Scalable Conformer Generation27 Nov
  3. 2023Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation29 Sep
  4. 2023Geometric Latent Diffusion Models for 3D Molecule Generation2 May
  5. 2023MiDi: Mixed Graph and 3D Denoising Diffusion for Molecule Generation17 Feb
  6. 2022DiGress: Discrete Denoising diffusion for graph generation29 Sep
  7. 2022MDM: Molecular Diffusion Model for 3D Molecule Generation13 Sep
  8. 2022Equivariant Diffusion for Molecule Generation in 3D31 Mar
  9. 2022GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation6 Mar

Other references · 73

  1. 2017Attention Is All You Needcited by 60,361
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  5. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  6. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  7. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  8. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  9. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  10. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,255
  11. 2023Visual Instruction Tuningcited by 7,869
  12. 2022Scalable Diffusion Models with Transformerscited by 6,054
38 more
  1. 2016Layer Normalizationcited by 5,669
  2. 2018How Powerful are Graph Neural Networks?cited by 4,393
  3. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,915
  4. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,761
  5. 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,598
  6. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
  7. 2021Variational Diffusion Modelscited by 1,153
  8. 2017MoleculeNet: A Benchmark for Molecular Machine Learningcited by 1,051
  9. 2019Strategies for Pre-training Graph Neural Networkscited by 971
  10. 2021E(n) Equivariant Graph Neural Networkscited by 965
  11. 2016Automatic chemical design using a data-driven continuous representation of moleculescited by 765
  12. 2018Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point cloudscited by 746
  13. 2022Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distributioncited by 739
  14. 2023PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallelcited by 660
  15. 2018Junction Tree Variational Autoencoder for Molecular Graph Generationcited by 652
  16. 2024TinyLlama: An Open-Source Small Language Modelcited by 618
  17. 2022GraphMAE: Self-Supervised Masked Graph Autoencoderscited by 448
  18. 2019Analyzing Learned Molecular Representations for Property Predictioncited by 375
  19. 2018Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generationcited by 364
  20. 2020GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generationcited by 305
  21. 2018Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Modelscited by 275
  22. 2022Torsional Diffusion for Molecular Conformer Generationcited by 263
  23. 2021Pre-training Molecular Graph Representation with 3D Geometrycited by 248
  24. 2020GEOM: Energy-annotated molecular conformations for property prediction and molecular generationcited by 243
  25. 2018Constrained Graph Variational Autoencoders for Molecule Designcited by 185
  26. 2021GraphDF: A Discrete Flow Model for Molecular Graph Generationcited by 180
  27. 2019Symmetry-adapted generation of 3d point sets for the targeted discovery of moleculescited by 172
  28. 2017Deep Reinforcement Learning for De-Novo Drug Designcited by 161
  29. 20213D Infomax improves GNNs for Molecular Property Predictioncited by 146
  30. 2023MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adaptercited by 128
  31. 2022ChemBERTa-2: Towards Chemical Foundation Modelscited by 127
  32. 2021MARS: Markov Molecular Sampling for Multi-objective Drug Discoverycited by 125
  33. 2021GeoMol: Torsional Geometric Generation of Molecular 3D Conformer Ensemblescited by 107
  34. 2021Inverse design of 3d molecular structures with conditional generative neural networkscited by 93
  35. 2018Learning Multimodal Graph-to-Graph Translation for Molecular Optimizationcited by 88
  36. 2019MolecularRNN: Generating realistic molecular graphs with optimized propertiescited by 82
  37. 2019Molecular Property Prediction: A Multilevel Quantum Interactions Modeling Perspectivecited by 73
  38. 2022Equivariant Energy-Guided SDE for Inverse Molecular Designcited by 73

50 most cited of 73 All 73 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 12; 2026: 5.

17 papers.

Sort:

By area:

  1. 2025GEOM-Drugs Revisited: Toward More Chemically Accurate Benchmarks for 3D Molecule Generationcited by 8
  2. 2025Language-Enhanced Representation Learning for Single-Cell Transcriptomicscited by 7
  3. 2025Sampling 3D Molecular Conformers with Diffusion Transformerscited by 6
  4. 2025DiffSpectra: Molecular Structure Elucidation from Spectra using Diffusion Modelscited by 5
  5. 2025Atomic Diffusion Models for Small Molecule Structure Elucidation from NMR Spectracited by 4
  6. 2026FlowMS: Flow Matching for De Novo Structure Elucidation from Mass Spectracited by 2
  7. 2025PRING: Rethinking Protein-Protein Interaction Prediction from Pairs to Graphscited by 1
  8. 2025Flow-Matching Based Refiner for Molecular Conformer Generationcited by 1
  9. 2025EnzyControl: Adding Functional and Substrate-Specific Control for Enzyme Backbone Generationcited by 1
  10. 2025MolSculpt: Sculpting 3D Molecular Geometries from Chemical Syntaxcited by 1
  11. 20253D-GSRD: 3D Molecular Graph Auto-Encoder with Selective Re-mask Decodingcited by 0
  12. 2025Learning 3D Anisotropic Noise Distributions Improves Molecular Force Field Modelingcited by 0
5 more
  1. 2025Chemistry-Enhanced Diffusion-Based Framework for Small-to-Large Molecular Conformation Generationcited by 0
  2. 2026SymDrift: One-Shot Generative Modeling under Symmetriescited by 0
  3. 2026Ligand-Conditioned Discrete Diffusion for Protein Sequence-Structure Co-Designcited by 0
  4. 2026Language-Informed Flow Matching for Trend-Guided Structure-Based 3D Molecular Generationcited by 0
  5. 2026TreeRef-BFN: Equivariance-Free De Novo Molecule Generation based on 2D Topology and Internal 3D Geometrycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Autoregressive latent diffusion for 3D molecule generation
  2. 2024Diffusion Models in textitDe Novo Drug Design
  3. 2025D3MES: Diffusion Transformer with multihead equivariant self-attention for 3D molecule generation
  4. 2023Learning Joint 2D & 3D Diffusion Models for Complete Molecule Generation
  5. 20243M-Diffusion: Latent Multi-Modal Diffusion for Language-Guided Molecular Structure Generation
  6. 2026MolHIT: Advancing Molecular-Graph Generation with Hierarchical Discrete Diffusion Models
  7. 2023Coarse-to-Fine: a Hierarchical Diffusion Model for Molecule Generation in 3D
  8. 2025MODA: A Unified 3D Diffusion Framework for Multi-Task Target-Aware Molecular Generation