Multi-step tool-calling on ACEBench
0.85End-to-End Success Rate
GPT-5
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.06 | 0.85 | 94.3 | |
2026.06 | 0.8 | 92 | |
2026.06 | 0.7 | 88.8 | |
2026.06 | 0.65 | 83.6 | |
2026.06 | 0.65 | 76.3 | |
2026.06 | 0.15 | 25.1 |
GPT-5
| Method | Links | ||
|---|---|---|---|
2026.06 | 0.85 | 94.3 | |
2026.06 | 0.8 | 92 | |
2026.06 | 0.7 | 88.8 | |
2026.06 | 0.65 | 83.6 | |
2026.06 | 0.65 | 76.3 | |
2026.06 | 0.15 | 25.1 |