A Fair Evaluation of Graph Foundation Models for Node Property Prediction

This study conducts a rigorous reevaluation of nine recent Graph Foundation Models (GFMs) for node property prediction, comparing them against strong Graph Neural Network (GNN) baselines to address the lack of unified evaluation standards in the field.

• The research evaluates nine recent GFMs designed for node property prediction tasks.
• These models are compared against strong Graph Neural Network (GNN) baselines.
• Only the most recent GFMs based on the Prior-data Fitted Networks paradigm outperform well-tuned GNNs.
• The superior GFMs achieve this performance at a higher inference cost.

The findings indicate that while specific newer GFMs can surpass traditional GNNs in predictive accuracy, they come with increased computational costs during inference.