Inductive Bio Integrates ADMET Prediction Models into Anthropic's Claude

The integration addresses a critical bottleneck in small molecule development: the complex balancing of potency against absorption, distribution, metabolism, excretion, and toxicity—collectively known as ADMET. Previously, these assessments required significant time and specialized resources. Inductive Bio’s models, which secured the top spot in the OpenADMET-ExpansionRx blind challenge against 370 industry and academic entries, are now available to streamline this evaluation.

By embedding these tools into Claude and Claude Science, researchers can reason about compound design rationale in real time without leaving their existing digital environment. According to Ben Birnbaum, co-founder and CTO of Inductive Bio, the move represents a shift in how scientists interact with predictive data, moving from disparate software silos to a fluid conversational workflow. To address industry privacy concerns, Inductive Bio confirmed that chemical structures processed through the connector are not retained or utilized to train the company's underlying AI models.