OGB-MolPCBA ogbg-molpcba Leaderboard
Molecular property prediction across 128 bioassay tasks (437k molecules); evaluated by Average Precision (AP) over the multi-task targets. · Metric: AP (higher is better) · 🟢 Updated 2h ago
| # | Model | AP | Paper |
|---|---|---|---|
| 1 | HyperFusion | 0.32 | — |
| 2 | HIG(pre-trained on PCQM4M) | 0.32 | — |
| 3 | Graphormer (pre-trained on PCQM4M) | 0.31 | — |
| 4 | 0.30 | — | |
| 5 | PAS | 0.30 | — |
| 6 | Nested GIN+virtual node (ensemble) | 0.30 | — |
| 7 | GINE+bot | 0.30 | — |
| 8 | CRaWl | 0.30 | — |
| 9 | GatedGCN+ | 0.30 | — |
| 10 | GINE+ w/ APPNP | 0.30 | — |
| 11 | PHC-GNN | 0.29 | — |
| 12 | GIN-AK | 0.29 | — |
| 13 | GINE+ w/ virtual nodes | 0.29 | — |
| 14 | DGN | 0.29 | — |
| 15 | RandomGIN-vn+FLAG | 0.29 | — |
| 16 | DeeperGCN+virtual node+FLAG | 0.28 | — |
| 17 | PNA | 0.28 | — |
| 18 | GIN+virtual node+FLAG | 0.28 | — |
| 19 | DeeperGCN+virtual node | 0.28 | — |
| 20 | GIN+virtual node | 0.27 | — |
| 21 | GCN+virtual node+FLAG | 0.25 | — |
| 22 | GCN+virtual node | 0.24 | — |
| 23 | GIN+FLAG | 0.24 | — |
| 24 | ChebNet | 0.23 | — |
| 25 | GIN | 0.23 | — |
| 26 | GCN+FLAG | 0.21 | — |
| 27 | MorganFP+Rand. Forest | 0.21 | — |
| 28 | GCN | 0.20 | — |