MPP Foundation Model Benchmark
Comprehensive benchmark from systematic survey of deep learning for molecular property prediction in the foundation model era. Provides unified evaluation across 4 paradigms (descriptor-based, GNN, pretrained, foundation models) with standardized splits and metrics.
Composite
50.3
Experimental validation
Retrospective
Stages
Lead ID / ADMETHit ID
Modalities
small molecule
Task types
classificationregression
Size
datasets: multiple standard (MoleculeNet, TDC subsets)
models_evaluated: 30
paradigms: 4
models_evaluated: 30
paradigms: 4
License
Other
First release
2026-04-17
Last updated
2026-04-17
Official site
Leaderboard
→ leaderboard
Dataset
→ dataset
Code / GitHub
→ repository
HuggingFace
→ HF
Paper
A Systematic Survey and Benchmark of Deep Learning for Molecular Property Prediction in the Foundation Model Era · Zongru Li, Xingsheng Chen, Honggang Wen, Regina Qianru Zhang, Ming Li, Xiaojin Zhang, Hongzhi Yin, Qiang Yang, Kwok-Yan Lam, Pietro Lio, Siu-Ming Yiu · 2026 · paper · doi:N/A — preprint · 0 citations
Flags
none
Experts
—
Groups
—
Hosted by
—
Related benchmarks
Rubric (7-criterion)
rigor
4
coverage
4
maintenance
3
adoption
1
quality
4
accessibility
2
industry_relevance
3
Notes
Valuable meta-benchmark tracing evolution from descriptors to foundation models. Includes industry perspective and highlights evaluation protocol challenges. Multi-institutional (HKU, NTU, Cambridge). Released April 2026, no code/leaderboard yet. Higher potential than current scores suggest once code is released.