ChemInsight: A Comprehensive Benchmark for Evaluating and Diagnosing Chemistry Understanding and Reasoning in Large Language Models

Changyu Hu, Yiming Zou, Xin Tang +7

We introduce ChemInsight, a benchmark designed to evaluate large language models (LLMs) on chemistry tasks that require multi-step reasoning and structured chemical understanding.…