OpenAI releases MedChemBench, a benchmark for medicinal chemistry
What happened
On November 20, OpenAI released MedChemBench, a medicinal chemistry benchmark that tests whether AI can read molecular structures, explain experimental results and predict compound properties, to see if it can help scientists screen drug candidates. The benchmark has 99 questions drawn from 117 papers and patents. OpenAI also published scores at the xhigh reasoning setting: GPT-6.2 Astra 84.5%, GPT-6 Astra 83.3%, GPT-6.1 Sol 82.0% and GPT-5.6 Sol 58.2%.
Written by AI from the coverage · updated 2 days ago
Coverage
Follow the reports to see the story from different sides.
- OpenAI NewsCan AI help scientists identify promising drug candidates?
OpenAI 发布 MedChemBench 基准,用 99 道源自 117 篇论文和专利的题目测试 AI 解读分子结构、解释实验结果并预测化合物性质的能力。在 xhigh 推理设置下,GPT‑6.2 Astra 获得 84.5% 的可用分数,GPT‑6 Astra 为 83.3%,GPT‑6.1 Sol 为 82.0%,GPT‑5.6 Sol 为 58.2%。
Heat over time
Heat now 3·Comparable peak 10(Nov 21 05:00)·Comparable change over 24 hours -49%
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.