High-Accuracy Quantum Chemistry Labels
Your molecules, labelled at CCSD(T) grade and tokenised as quantum inputs for your model. Faster to make, and your model trains on far less data.
Within chemical accuracy of CCSD(T)/CBS-level reference labels.
Faster than the standard CCSD(T)-level label recipe.
Fewer labels than a standard descriptor model for the same error, on the same CCSD(T)/CBS-level label and held-out molecules. Chemical accuracy where the descriptor model stops at 1.8 kcal/mol.
Built for drug discovery teams
AI drug discovery teams
Train your models on gold-standard data and rank compounds you can trust, with far less data.
Computational chemistry
Gold-standard numbers for the molecules where DFT and force fields go wrong.
Discovery programmes
Answers in days, not months, so your design cycle never waits on quantum chemistry.
Already in researchers' hands
Fewer labelled molecules than a standard model, on our FCI-label benchmark. Published on ChemRxiv.
Read the paper →Methods we support
CCSD(T)
Gold-standard accuracy for small and drug-fragment-sized organic molecules.
FCI
Exact reference data for small systems.
DMRG
Strongly correlated systems, such as metal centres. On request.
ANI-1ccx / ANI-2x
Drop-in format for your existing training pipeline.
Your molecules. Gold-standard labels. In days.
Send us a molecule set. We return it labelled and tokenised, with a report against the labels you use today.
Request a pilot