Tool-augmented agent execution on MCP-Bench
46.8Task Fulfillment
ReAct
Updated 22d ago
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2026.05 | 46.8 | 36.6 | 41.5 | 31.2 | 40.5 | 19.8 | 77.7 | 0 | 41.7 | 36.4 | 30.1 | 38.9 | 36.8 | |
2026.05 | 38.5 | 38.7 | 42.5 | 47.8 | 36.9 | 20.6 | 100 | 75.4 | 38.6 | 45.2 | 28.8 | 87.7 | 50 | |
2026.05 | 34.5 | 47.5 | 55.3 | 58.4 | 33.3 | 27.5 | 100 | 79.9 | 41 | 56.9 | 30.4 | 89.9 | 54.6 | |
2026.05 | 33.7 | 42.6 | 45.2 | 40.4 | 29.6 | 26.8 | 100 | 85.8 | 38.2 | 42.8 | 28.2 | 92.9 | 50.5 | |
2026.05 | 32.8 | 51 | 59.6 | 61 | 33.6 | 30.9 | 100 | 82.8 | 41.9 | 60.3 | 32.2 | 91.4 | 56.5 | |
2026.05 | 31.6 | 26.6 | 33.6 | 48.4 | 20.5 | 20.2 | 100 | 100 | 29.1 | 41 | 20.3 | 100 | 47.6 | |
2026.05 | 31.4 | 47.2 | 53.1 | 65.9 | 27.1 | 29.9 | 100 | 78.4 | 39.3 | 59.5 | 28.5 | 89.2 | 54.1 |