Tool-Use Agent Evaluation on BFCL Multiturn (OOD) v3 (test)
48Base Rate
TopoCurate-RL
Updated 1mo ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.03 | 48 | 29.5 | 22.5 | 32.5 | |
2026.03 | 45 | 24.5 | 16.5 | 26 | |
2026.03 | 43 | 26.5 | 13.5 | 29.5 | |
2026.03 | 42 | 21.5 | 20 | 30 | |
| 41 | 23.5 | 12.5 | 24.5 | ||
2026.03 | 37.5 | 21.5 | 17.5 | 20.5 | |
2026.03 | 35 | 19.5 | 21 | 28.5 | |
2026.03 | 34 | 21 | 15.5 | 24.5 | |
2026.03 | 29.5 | 16 | 13.5 | 18 | |
| 26.5 | 13.5 | 7.5 | 15 | ||
2026.03 | 26.5 | 12.5 | 9 | 16 | |
2026.03 | 25.5 | 14.5 | 10.5 | 16 | |
2026.03 | 24.5 | 14 | 5 | 17 | |
2026.03 | 24 | 10 | 11 | 16.5 | |
2026.03 | 23 | 13.5 | 4.5 | 13.5 | |
2026.03 | 21 | 11 | 11.5 | 15 | |
2026.03 | 4.5 | 2 | 0.5 | 8 | |
2026.03 | 4 | 2.5 | 1.5 | 7.5 |