CoreX is built for scientific scrutiny: defined use cases, documented protocols, quality-controlled biology, auditable data systems, and transparent limitations. Biological intelligence is only valuable if it can be inspected, questioned, reproduced, and improved.
CoreX deploys the Quris-AI Bio-AI platform for sovereign, population-specific medicine programs. In defined drug-induced liver injury (DILI) benchmarks and blinded prospective studies, the platform has demonstrated performance against conventional biology-only and AI-only approaches. Full study designs, compound sets, comparator methods, confidence intervals, and limitations are available to qualified partners under NDA.
Platform evidence shown in defined DILI and blinded prospective study contexts. Not a general claim across all diseases, medicines, organs, or populations.
Liver-toxicity (DILI) prediction in a retrospective benchmark, versus 80% for 2D biology and 76% for 3D biology alone.
Quris-AI platform evidenceOn the same DILI benchmark, versus 25% for 2D biology and 50% for 3D biology alone.
Quris-AI platform evidenceIn independent blinded prospective studies, versus 68% for AI alone and 63% for biology alone.
Quris-AI platform evidenceIn one program, the same platform flagged a safety signal in a major drug candidate that standard preclinical models did not catch.
"This technology flags drug toxicity that all our models completely missed."CAMS 2024 · blinded validation study
Evaluated in independent, blinded prospective studies with leading global pharmaceutical partners.
Metrics reflect defined contexts of use, such as liver-toxicity (DILI) prediction. Full study designs, compound sets, comparator methods, confidence intervals, and limitations are available to qualified partners under NDA.
Each result is shown with its context of use, study type, comparator, and its limitation. Evidence is only useful if you can see its edges.
A model is never validated in general. It is validated for a defined context of use: a specific question, under specific conditions, with a stated boundary.
Living donor-specific models.
Controlled organ-on-chip and organoid response systems.
Structured biological response signals.
Governed AI trained on biological response.
Partner insight for defined use cases.
Compounding evidence across cohorts.
Every claim is tied to a specific scientific or development use case.
Clear endpoints, inclusion criteria, controls, and analysis plans.
Documented characterization, acceptance criteria, and batch controls.
Data provenance, version control, leakage prevention, and auditability.
| Capability | Status | Public wording |
|---|---|---|
| Drug toxicity modeling | Qualified programs | Available for qualified partner programs |
| Pharmacokinetic modeling | Defined programs | Available under defined programs |
| Population cohort modeling | Governed programs | Available through governed population programs |
| Multi-organ response | Research-stage | Research-stage / partner development |
| Regenerative applications | Future research | Not offered as a clinical therapy |
Capability labels are intentionally high-level. Technical validation materials, study design, and context-specific performance data are available to qualified partners under appropriate confidentiality frameworks.