1910 combines molecular data, AI models, and laboratory automation within its ITO drug-discovery platform. Its published CANDID-CNS work uses an attentive graph neural network to predict blood-brain-barrier penetration of complex small molecules and to distinguish stereoisomers.
In a December 2025 company announcement of the peer-reviewed study, CANDID-CNS achieved 87% area under the precision-recall curve on its beyond-rule-of-five test set, compared with 56% for Pfizer's CNS MPO score. The comparison measures this specific prediction task, while 1910's partnering page shows additional co-engineering and co-discovery uses for its platform.
1910 is commercializing AI-assisted drug discovery through several partnership models. Its current partnering page lists customized ITO infrastructure, co-engineering, co-discovery, asset licensing, and data licensing as available routes for pharmaceutical and biotechnology organizations.
The company's December 2025 CANDID-CNS publication describes a graph-neural-network model for predicting brain penetration in complex small molecules, with results reported against Pfizer's CNS MPO score. The combination of scientific publication and available collaboration models gives partners concrete ways to evaluate its technology for CNS and other drug-discovery programs.
