Molecular property prediction on BBBP
0.968ROC AUCDMMRL
Evaluation Results
| Method | Links | |
|---|---|---|
| DMMRL2026.03 | 0.968 | |
| SGGRL2026.03 | 0.967 | |
| MvMRL2026.03 | 0.962 | |
| GROVER2026.03 | 0.955 | |
| Transformer2026.03 | 0.944 | |
| GraSeq2026.03 | 0.942 | |
| FP-GNN2026.03 | 0.935 | |
| N-Gram2026.03 | 0.912 | |
| MPNN2026.03 | 0.91 | |
| RNN2026.03 | 0.902 | |
| TrimNet2026.03 | 0.889 | |
| ChemATP2025.12 | 0.8731 | |
| MGCNModel Type=Graph-based2026.06 | 0.85 | |
| SchNetModel Type=Graph-based2026.06 | 0.848 | |
| GROVERModel Type=Graph-based2026.06 | 0.841 | |
| Qwen34B-InstStrategy=+MOLE-RAG2026.06 | 0.801 | |
| ChemDFM14BStrategy=Baseline2026.06 | 0.79 | |
| ChemDFM14BStrategy=+MOLE-RAG2026.06 | 0.781 | |
| GPT-4oStrategy=+MOLE-RAG2026.06 | 0.77 | |
| LLM4SD2025.12 | 0.756 | |
| GPT-4o-miniStrategy=+MOLE-RAG2026.06 | 0.747 | |
| Mistral7B-InstStrategy=+MOLE-RAG2026.06 | 0.746 | |
| CHCLModel=CHCL2026.04 | 0.7433 | |
| GMCLModel=GMCL2026.04 | 0.738 | |
| MolCLRGINModel Type=Graph-based2026.06 | 0.735 | |
| AutoGCLModel=AutoGCL2026.04 | 0.7336 | |
| MolCLRGCNModel Type=Graph-based2026.06 | 0.732 | |
| UniMol2025.12 | 0.729 | |
| GraphLoGPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.725 | |
| GraphMVP2025.12 | 0.724 | |
| MolCLRCMPNNModel Type=Graph-based2026.06 | 0.724 | |
| GraphMAEPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.72 | |
| MoMu2025.12 | 0.705 | |
| JOAOModel=JOAO2026.04 | 0.7022 | |
| JOAOPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.702 | |
| GraphormerType=Graph-based / Full Supervision2026.07 | 0.702 | |
| AD-GCLModel=AD-GCL2026.04 | 0.7001 | |
| InstructMol2025.12 | 0.7 | |
| GREAEncoder=GCN [12]2022.06 | 0.6986 | |
| GraphCLPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.697 | |
| GREAEncoder=GIN [37]2022.06 | 0.697 | |
| GraphCLModel=GraphCL2026.04 | 0.6968 | |
| GCN2026.03 | 0.69 | |
| IRMEncoder=GCN [12]2022.06 | 0.6892 | |
| InfomaxPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.688 | |
| InfomaxModel=Infomax2026.04 | 0.688 | |
| IRMEncoder=GIN [37]2022.06 | 0.6835 | |
| DIR+REPAUGEncoder=GIN [37]2022.06 | 0.6813 | |
| U-NETSPOOLEncoder=GIN [37]2022.06 | 0.6809 | |
| ContextPredModel=ContextPred2026.04 | 0.68 | |
| OOD-GNNEncoder=GCN [12]2022.06 | 0.6795 | |
| GREA-REPAUGEncoder=GIN [37]2022.06 | 0.6737 | |
| GREA-REPAUGEncoder=GCN [12]2022.06 | 0.6736 | |
| EdgePredModel=EdgePred2026.04 | 0.673 | |
| OOD-GNNEncoder=GIN [37]2022.06 | 0.671 | |
| Weave2026.03 | 0.671 | |
| U-NETSPOOLEncoder=GCN [12]2022.06 | 0.6709 | |
| STABLEGNNEncoder=GIN [37]2022.06 | 0.6695 | |
| GATType=Graph-based / Full Supervision2026.07 | 0.665 | |
| DIR+REPAUGEncoder=GCN [12]2022.06 | 0.663 | |
| SELFATTNPOOLEncoder=GIN [37]2022.06 | 0.6624 | |
| SELFATTNPOOLEncoder=GCN [12]2022.06 | 0.6602 | |
| No PretrainModel=No Pretrain2026.04 | 0.658 | |
| GINType=Graph-based / Full Supervision2026.07 | 0.658 | |
| DIREncoder=GCN [12]2022.06 | 0.6559 | |
| STABLEGNNEncoder=GCN [12]2022.06 | 0.6552 | |
| No-pretrainPre-training dataset=None, Protocol=Supervised2022.05 | 0.655 | |
| GCNType=Graph-based / Full Supervision2026.07 | 0.649 | |
| ContextPredPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.643 | |
| AttrMaskingPre-training dataset=ZINC15, Protocol=Fine-tuning2022.05 | 0.643 | |
| AttrMaskingModel=AttrMasking2026.04 | 0.643 | |
| MolBasicShot=4-shot2026.07 | 0.633 | |
| MolRAG2025.12 | 0.6291 | |
| CoTPrompting Strategy=Chain-of-Thought2025.12 | 0.6121 | |
| MolBasicShot=2-shot2026.07 | 0.611 | |
| KVPLMType=Pre-training2026.07 | 0.602 | |
| GIMLETType=Pre-training2026.07 | 0.594 | |
| DIREncoder=GIN [37]2022.06 | 0.5843 | |
| MolRAGShot=4-shot2026.07 | 0.572 | |
| MolRAGShot=2-shot2026.07 | 0.564 | |
| MolBasicShot=1-shot2026.07 | 0.55 | |
| GPT-4o-miniStrategy=Baseline2026.06 | 0.549 | |
| MolRAGShot=0-shot2026.07 | 0.546 | |
| MolRAGShot=1-shot2026.07 | 0.546 | |
| Galactica-1.3BType=Pre-training2026.07 | 0.539 | |
| GPT-4oStrategy=Baseline2026.06 | 0.534 | |
| Qwen34B-InstStrategy=Baseline2026.06 | 0.53 | |
| MolBasicShot=0-shot2026.07 | 0.514 | |
| MoMuType=Pre-training2026.07 | 0.498 | |
| DAPrompting Strategy=Direct Answer2025.12 | 0.4918 | |
| Llama3.2-3BStrategy=+MOLE-RAG2026.06 | 0.483 | |
| Llama3.2-3BStrategy=Baseline2026.06 | 0.482 | |
| Mistral7B-InstStrategy=Baseline2026.06 | 0.461 |