List your company

Text-to-SQL on Spider2-Lite SQLite v1.0 (test)

64.44Pass@1 AccuracyREFORCE
35.94443.34250.7458.138May 20, 2026

Evaluation results

MethodLinks
REFORCEBackbone=GPT-5.4, K (number of candidates)=dynamic2026.0564.44—58.52
DIVSKILL-SQLBackbone=GPT-5.4, K (number of candidates)=82026.0562.8784.4471.85
CHASE-SQLBackbone=GPT-5.4, K (number of candidates)=82026.0561.5783.766.67
DIN-SQLBackbone=GPT-5.42026.0537.04——