Deep Dive into TableRecordMatch a New Metric for Evaluating Parsing Accuracy on Complex Tables

ParseBench introduces 5 novel metrics for measuring document OCR accuracy for AI agents. For tables specifically, we introduce TableRecordMatch — a metric that evaluates tables the way downstream systems actually consume them: as records where each row is a set of values keyed by column headers. Column reordering? No penalty. Transposed headers or dropped columns? Heavy penalty. Read more about ParseBench here: https://www.llamaindex.ai/blog/parsebench?utm_medium=video&utm_source=youtube&utm_campaign=2026-apr-