Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
CodeXGLUEmicrosoft.github.io · Low evidenceCodeXGLUE is a benchmark for general language understanding evaluation of code, comprising 14 datasets across 10 programming tasks. It provides three baseline models—CodeBERT, CodeGPT, and an Encoder-Decoder framework—to support tasks like code clone detection, defect detection, and code translation.
Parse Score