Chemistry

MedChemBench (Internal)

OpenAI's own drug-discovery test: reading chemical structures, predicting how potent or toxic a compound will be, choosing between candidate molecules, and planning a synthesis route — the everyday judgement calls of a medicinal chemist. Higher is better.

Rankings

Higher is better

MedChemBench (Internal) — frequently asked questions

What is MedChemBench (Internal)?
OpenAI's own drug-discovery test: reading chemical structures, predicting how potent or toxic a compound will be, choosing between candidate molecules, and planning a synthesis route — the everyday judgement calls of a medicinal chemist. Higher is better.
Which AI model scores highest on MedChemBench (Internal)?
GPT-6 Astra by OpenAI holds the best MedChemBench (Internal) result among tracked models, at 49.7% (released Sep 3 2026). Higher scores are better on this benchmark.
What are the top 1 models on MedChemBench (Internal)?
1. GPT-6 Astra (OpenAI) — 49.7%.
How many models have a published MedChemBench (Internal) score?
1 tracked model has a published MedChemBench (Internal) score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.