Insilico Medicine announced on July 30 the launch of the Drug Discovery and Development Benchmarks as a Service, a standardized evaluation framework designed to validate artificial intelligence models used in drug discovery. The company said the new system aims to address concerns about data contamination in existing benchmarks, where AI models may achieve high scores by memorizing test questions rather than demonstrating real-world capability.
The Drug Discovery and Development Benchmark offers two complementary evaluation suites. The first, Drug Discovery Foundations, includes over 300 evaluations using proprietary out-of-distribution test sets and decontaminated public data to measure core competencies such as disease biology, molecular property prediction and optimization, retrosynthesis, structure-based design, and clinical development. The second suite, Drug Candidate Essentials, assesses a model's ability to handle an end-to-end drug discovery program from hit identification through preclinical candidate nomination.
According to Insilico Medicine, benchmarking is accessible for any organization developing or leveraging frontier AI for drug discovery. Models can be evaluated via a standard chat-completions API by contacting the company directly. Results are scored against expert reference baselines and delivered with a verified score report that organizations can use internally or share with partners. Those seeking public recognition may also publish their results on the DDD Benchmark public leaderboard.
"As AI systems increasingly act as agents (planning experiments, reasoning over experimental data, and calling MCP-compatible tools), the DDD Benchmark is designed to measure whether those capabilities translate into sound, real-world drug discovery decisions rather than strong performance on abstract tasks," Insilico Medicine said.
The new benchmark builds on Insilico's Pharma.AI platform and MMAI Gym post-training environment for scientific AI. Since its founding, Insilico has nominated 31 preclinical candidates and received more than 10 investigational new drug clearances while reducing the timeline for preclinical candidate nomination compared with traditional methods. Its lead program is Rentosertib (ISM001-055), an AI-discovered TNIK inhibitor currently in Phase III development.
