Papers › Long Range Graph Benchmark

Long Range Graph Benchmark

16 Jun 2022arXiv:2206.08164archive 2025-07-28

Vijay Prakash Dwivedi, Ladislav Rampášek, Mikhail Galkin, Ali Parviz, Guy Wolf, Anh Tuan Luu, Dominique Beaini

Graph Neural Networks (GNNs) that are based on the message passing (MP) paradigm generally exchange information between 1-hop neighbors to build node representations at each layer. In principle, such networks are not able to capture long-range interactions (LRI) that may be desired or necessary for learning a given task on graphs. Recently, there has been an increasing interest in development of Transformer-based methods for graphs that can consider full node connectivity beyond the original sparse structure, thus enabling the modeling of LRI. However, MP-GNNs that simply rely on 1-hop message passing often fare better in several existing graph benchmarks when combined with positional feature representations, among other innovations, hence limiting the perceived utility and ranking of Transformer-like architectures. Here, we present the Long Range Graph Benchmark (LRGB) with 5 graph learning datasets: PascalVOC-SP, COCO-SP, PCQM-Contact, Peptides-func and Peptides-struct that arguably require LRI reasoning to achieve strong performance in a given task. We benchmark both baseline GNNs and Graph Transformer networks to verify that the models which capture long-range dependencies perform significantly better on these tasks. Therefore, these datasets are suitable for benchmarking and exploration of MP-GNNs and Graph Transformer architectures that are intended to capture LRI.

PaperPDFCode

In Syntology View this paper on Syntology, its page in Syntology's graph. That page lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

vijaydwivedi75/lrgb officialmentioned in papermentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Multiresolution Graph Transformers and Wavelet Positional Encoding for Learning... (arXiv:2302.08647; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
zml72062/dr-fwl-2 mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Distance-Restricted Folklore Weisfeiler-Leman GNNs with Provable Cycle Counting Power (arXiv:2309.04941; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

BenchmarkingGraph ClassificationGraph LearningGraph RegressionLink PredictionNode Classification

Datasets

Introduced by this paper, per the archive.

Long Range Graph Benchmark (LRGB)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Graph Classification Peptides-func SAN+RWSE AP 0.6439±0.0075 #34 of 44 Archive leaderboard report
Graph Classification Peptides-func SAN+LapPE AP 0.6384±0.0121 #36 of 44 Archive leaderboard report
Graph Classification Peptides-func Transformer+LapPE AP 0.6326±0.0126 #37 of 44 Archive leaderboard report
Graph Classification Peptides-func GatedGCN+RWSE AP 0.6069±0.0035 #38 of 44 Archive leaderboard report
Graph Classification Peptides-func GCN AP 0.5930±0.0023 #41 of 44 Archive leaderboard report
Graph Classification Peptides-func GatedGCN AP 0.5864±0.0077 #42 of 44 Archive leaderboard report
Graph Classification Peptides-func GINE AP 0.5498±0.0079 #44 of 44 Archive leaderboard report
Graph Regression Peptides-struct Transformer+LapPE MAE 0.2529±0.0016 #26 of 39 Archive leaderboard report
Graph Regression Peptides-struct SAN+RWSE MAE 0.2545±0.0012 #28 of 39 Archive leaderboard report
Graph Regression Peptides-struct SAN+LapPE MAE 0.2683±0.0043 #33 of 39 Archive leaderboard report
Graph Regression Peptides-struct GatedGCN+RWSE MAE 0.3357±0.0006 #35 of 39 Archive leaderboard report
Graph Regression Peptides-struct GatedGCN MAE 0.3420±0.0013 #36 of 39 Archive leaderboard report
Graph Regression Peptides-struct GCN MAE 0.3496±0.0013 #38 of 39 Archive leaderboard report
Graph Regression Peptides-struct GINE MAE 0.3547±0.0045 #39 of 39 Archive leaderboard report
Link Prediction PCQM-Contact SAN+LapPE Hits@1 0.1355±0.0017 #3 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+LapPE Hits@10 0.8478±0.0044 #3 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+LapPE Hits@3 0.4004±0.0021 #3 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+LapPE MRR 0.3350±0.0003 #3 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GINE Hits@1 0.1337±0.0013 #4 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GINE Hits@10 0.8147±0.0062 #4 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GINE Hits@3 0.3642±0.0043 #4 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GINE MRR 0.3180±0.0027 #4 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GCN Hits@1 0.1321±0.0007 #6 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GCN Hits@10 0.8256±0.0006 #6 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GCN Hits@3 0.3791±0.0004 #6 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GCN MRR 0.3234±0.0006 #6 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+RWSE Hits@1 0.1312±0.0016 #7 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+RWSE Hits@10 0.8550±0.0024 #7 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+RWSE Hits@3 0.4030±0.0008 #7 of 18 Archive leaderboard report
Link Prediction PCQM-Contact SAN+RWSE MRR 0.3341±0.0006 #7 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN+RWSE Hits@1 0.1288±0.0013 #8 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN+RWSE Hits@10 0.8517±0.0005 #8 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN+RWSE Hits@3 0.3808±0.0006 #8 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN+RWSE MRR 0.3242±0.0008 #8 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN Hits@1 0.1279±0.0018 #9 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN Hits@10 0.8433±0.0011 #9 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN Hits@3 0.3783±0.0004 #9 of 18 Archive leaderboard report
Link Prediction PCQM-Contact GatedGCN MRR 0.3218±0.0011 #9 of 18 Archive leaderboard report
Link Prediction PCQM-Contact Transformer+LapPE Hits@1 0.1221±0.0011 #10 of 18 Archive leaderboard report
Link Prediction PCQM-Contact Transformer+LapPE Hits@10 0.8517±0.0039 #10 of 18 Archive leaderboard report
Link Prediction PCQM-Contact Transformer+LapPE Hits@3 0.3679±0.0033 #10 of 18 Archive leaderboard report
Link Prediction PCQM-Contact Transformer+LapPE MRR 0.3174±0.0020 #10 of 18 Archive leaderboard report
Node Classification COCO-SP GatedGCN macro F1 0.2641±0.0045 #9 of 19 Archive leaderboard report
Node Classification COCO-SP Transformer+LapPE macro F1 0.2618±0.0031 #10 of 19 Archive leaderboard report
Node Classification COCO-SP SAN+LapPE macro F1 0.2592±0.0158 #11 of 19 Archive leaderboard report
Node Classification COCO-SP GatedGCN+LapPE macro F1 0.2574±0.0034 #12 of 19 Archive leaderboard report
Node Classification COCO-SP SAN+RWSE macro F1 0.2434±0.0156 #13 of 19 Archive leaderboard report
Node Classification COCO-SP GINE macro F1 0.1339±0.0044 #17 of 19 Archive leaderboard report
Node Classification COCO-SP GCN macro F1 0.0841±0.0010 #19 of 19 Archive leaderboard report
Node Classification PascalVOC-SP SAN+LapPE macro F1 0.3230±0.0039 #10 of 21 Archive leaderboard report
Node Classification PascalVOC-SP SAN+RWSE macro F1 0.3216±0.0027 #11 of 21 Archive leaderboard report
Node Classification PascalVOC-SP GatedGCN macro F1 0.2873±0.0219 #14 of 21 Archive leaderboard report
Node Classification PascalVOC-SP GatedGCN+LapPE macro F1 0.2860±0.0085 #15 of 21 Archive leaderboard report
Node Classification PascalVOC-SP Transformer+LapPE macro F1 0.2694±0.0098 #17 of 21 Archive leaderboard report
Node Classification PascalVOC-SP GCN macro F1 0.1268±0.0060 #20 of 21 Archive leaderboard report
Node Classification PascalVOC-SP GINE macro F1 0.1265±0.0076 #21 of 21 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutGraph TransformerLabel SmoothingLapEigenLaplacian PELayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections