Multi-turn tool-calling on ACEBench
15End-to-End Success Rate
GPT-5
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.06 | 15 | 77.8 | |
2026.06 | 11 | 70.3 | |
2026.06 | 9 | 69.4 | |
2026.06 | 9 | 69.5 | |
2026.06 | 8 | 61.2 | |
2026.06 | 4 | 36.2 |
GPT-5
| Method | Links | ||
|---|---|---|---|
2026.06 | 15 | 77.8 | |
2026.06 | 11 | 70.3 | |
2026.06 | 9 | 69.4 | |
2026.06 | 9 | 69.5 | |
2026.06 | 8 | 61.2 | |
2026.06 | 4 | 36.2 |