pepADMET
First comprehensive AI platform for peptide ADMET prediction. 36,643 high-quality data entries from open databases, covering 19-29 ADMET endpoints. Supports linear, cyclic, natural, and modified peptides across multiple species.
Composite
80.0
Experimental validation
Retrospective
Stages
IND-enablingLead ID / ADMET
Modalities
peptidesmall-molecule
Task types
admet_predictionproperty_prediction
Size
entries: 643
datasets: 26
-29_endpoints: 19
datasets: 26
-29_endpoints: 19
License
Academic
First release
2026-01
Last updated
2026-01
Official site
Leaderboard
→ leaderboard
Dataset
→ dataset
Code / GitHub
→ repository
HuggingFace
→ HF
Paper
pepADMET: A Comprehensive Platform for Peptide ADMET Prediction · · 2026 · paper · 3 citations
Flags
biologicspeptide_therapeuticsfirst_in_class
Experts
—
Groups
—
Hosted by
—
Related benchmarks
Rubric (7-criterion)
rigor
4
coverage
5
maintenance
3
adoption
3
quality
4
accessibility
4
industry_relevance
5
Notes
Fills critical gap in peptide ADMET prediction (previous tools focused on small molecules). Covers Caco-2, PAMPA, BBB, half-life, toxicity. Supports modified peptides unlike most tools. Chinese research group.