1910 offers Candid-Cns
ProductOperating Status: Currently Offered
CANDID-CNS is 1910's AI model for predicting small-molecule blood-brain barrier penetration in beyond-Rule-of-5 chemical space.
CANDID-CNS is 1910's AI model for predicting small-molecule blood-brain barrier penetration in beyond-Rule-of-5 chemical space.
CANDID-CNS publishes benchmark metrics for its stated prediction tasks: 90% versus 68% AUPRC for Rule-of-5 molecules, 87% versus 55% for beyond-Rule-of-5 molecules, and 71% versus 60% AUROC for stereoisomer distinction against the named CNS MPO comparator. These are the product page's declared comparisons.
The offering pairs the predictive model with an access workflow and 1910 co-engineering. The page says partners can collaborate on data, AI models, optimized compute, and interactive tools, extending the model into bespoke discovery modules rather than limiting it to a single fixed prediction screen.
CANDID-CNS is focused on prediction of blood-brain-barrier and related CNS properties, so it does not by itself establish broader efficacy, safety, or clinical outcomes for a molecule. The published metrics measure model tasks and do not substitute for downstream experimental or clinical validation.
Access is controlled through an application form or login, and the page describes co-engineering as a separate collaboration path. Those access requirements may add qualification and integration work for organizations seeking to use the model in an existing discovery workflow.
CANDID-CNS is positioned for small-molecule CNS discovery, especially compounds in beyond-Rule-of-5 chemical space where blood-brain-barrier prediction is a stated problem. The official page connects the model to CNS therapeutics and offers direct access plus a partner co-engineering route.
1910's page also describes bespoke modules built with partner data, models, compute, and interactive tools. That supports a workflow and collaboration-led market direction; the source does not state market size, customer count, or revenue.
CANDID-CNS is 1910's AI model for predicting blood-brain-barrier penetration of small molecules, including compounds beyond the traditional Rule-of-5 space. The official page highlights stereochemistry and CNS therapeutics as the product context and provides an application form and login for access.
The page describes core in-silico modules for Rule-of-5 and beyond-Rule-of-5 molecules plus stereoisomer distinction. It also presents a co-engineering service for bespoke small- or large-molecule discovery modules that can be used within a partner organization.