Agent tool-use planning on EarthBench 104 Tools (test)
69.8Tool Selection Accuracy (TSA)
Qwen3-235B
Updated 27d ago
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
2026.05 | 69.8 | 65.5 | 60.4 | 32.6 | 73.6 | 61.2 | 42.9 | |
2026.05 | 64.3 | 62.1 | 52.3 | 35.5 | 71.5 | 55.3 | 30.8 | |
2026.05 | 62.9 | 59.2 | 53 | 17.4 | 53.1 | 44 | 26.1 | |
2026.05 | 54.1 | 50.5 | 41 | 56.5 | 77 | 64.9 | 52.6 | |
2026.05 | 53.7 | 47.9 | 42.1 | 26.1 | 61.8 | 45.8 | 18.3 | |
2026.05 | 51.1 | 48.8 | 43.2 | 60.9 | 78.4 | 66.3 | 50 | |
2026.05 | 48.9 | 44.7 | 34.7 | 37 | 66.8 | 51.8 | 6.6 | |
2026.05 | 39.9 | 36.9 | 27.9 | 6.5 | 44.8 | 35.8 | 3.4 |