Medscape reported that its Clarus model scored 88.6% on NOHARM v2 and 85.0% on HealthBench Professional, benchmarks for clinical AI safety and performance. On NOHARM v2, which examines the frequency and severity of potentially harmful errors in AI-generated medical recommendations, Medscape said its 88.6% score exceeds the three highest reported competitor scores of 86.2%, 84.5% and 80.0%. On HealthBench Professional, which assesses AI responses to health-related scenarios against physician-defined standards for answer quality, Medscape said its 85.0% score surpasses the three highest published competitor scores of 69.2%, 65.6% and 65.3%. Joshua Conrad, PharmD, Group VP Editorial & Content Strategy at Medscape, said the results reflect rapid progress in developing AI for healthcare and a focus on systematic testing and continued refinement. Medscape also shared details of ClinVal, its proprietary clinician-developed evaluation framework, which assesses AI-generated answers against an evidence-based set of 16 criteria across dimensions including safety, accuracy, relevance, and traceability, established by Medscape clinicians and refined with AI engineering teams. The process combines automated AI assessments with clinician review, and findings help Medscape identify opportunities to enhance the product, with answers re-evaluated following changes to assess their impact.