The first comprehensive benchmark for evaluating cross-system SQL translation systems
Leaderboard • Documentation • Submit Results • Paper
09/2025: Our paper "PARROT: A Benchmark for Evaluating LLMs in Cross-System SQL Translation" has been accepted by NeurIPS 2025! :tada: :tada: :tada:
05/2025: We have released PARROT-1.0 (28,003 translation pairs from 38 open-source… See the full description on the dataset page:
https://huggingface.co/datasets/weizhoudb/PARROT.