Tag
A benchmark comparing AI models ranging from 2B to 35B parameters on a challenging task of extracting structured data from HTML, evaluating their performance and accuracy.