Text-to-SQL on Spider 120 2018/2025 (test)
88.3SRRAPEX MAS
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| APEX MASBase Model=Sonnet2026.04 | 88.3 | 97.1 | 36.5 | 47 | 67.5 | |
| Adaptive SkillTime (s)=170, Cost ($)=0.872026.04 | 5.9 | 47.7 | 85 | 57.9 | 65.8 | |
| MAS CompilerTime (s)=222, Cost ($)=0.852026.04 | 1.7 | 10.3 | 22.6 | 13.4 | 57.5 | |
| Claude Code RawTime (s)=103, Cost ($)=0.462026.04 | 0.8 | 8.4 | 21.7 | 11.4 | 60 |