Benchmark RadarSC3
SC3 is a multi-solvent solubility benchmark built on BigSolDB v2.1 with 101,535 measurements over 1,327 solutes and 206 solvents. It features a reproducible curation pipeline, nested Gold/Silver/Bronze consensus tiers, leakage-checked splits, and a multi-solvent metric suite (PS-RMSE, Z-RMSE).
- Released
- 2026-06-03
- Readiness
- Paper only
- Primary field
- Science & Research
Why it matters
Existing solubility benchmarks vary in curation and evaluation, hiding model failures. SC3 provides a standardized, reproducible benchmark with calibrated aleatoric limits to better measure model performance.
Motivation
Solubility prediction is a standard benchmark in computational chemistry, yet multi-solvent models which reportedly approach the experimental-noise ceiling (i.e.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.