Retrieval on retrieval benchmark (hold-out)
84Hit Rate @ 5GPT-5.4
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-5.4Turns=M, Tool Setup=Web Search + Toolset2026.07 | 84 | 87 | 89.5 | 56.8 | 71.8 | 41.6 | 43.8 | 5.6 | 1 | 0.1508 | |
| Qwen3.5-397BTurns=M, Tool Setup=Workflow2026.07 | 82 | 85 | 86.5 | 42.1 | 67.9 | 35.5 | 36 | 12.5 | 1.22 | 0.0657 | |
| GPT-5.4Turns=1, Tool Setup=Web Search + Toolset2026.07 | 79.5 | 82 | 84.5 | 53.5 | 62.8 | 37.1 | 39.4 | 2.6 | 1.07 | 0.3115 | |
| GPT-5.4Turns=M, Tool Setup=Web Search2026.07 | 79 | 85 | 89.5 | 48.2 | 65.3 | 35.3 | 37.5 | 5.6 | 1 | 0.1404 | |
| Qwen3.5-397BTurns=M, Tool Setup=Toolset2026.07 | 78.5 | 88.5 | 91.5 | 48.6 | 62.1 | 38 | 41.6 | 7.3 | 1.07 | 0.1082 | |
| PAPERPILOT-9BTurns=M, Tool Setup=Workflow + Toolset2026.07 | 77 | 83.5 | 89.5 | 40 | 59.4 | 32.5 | 35.4 | 42.8 | 1.66 | 0.018 | |
| Qwen3.5-397BTurns=1, Tool Setup=Toolset2026.07 | 74 | 87 | 89.5 | 43.2 | 57.8 | 34.1 | 37 | 15.7 | 1.67 | 0.047 | |
| GPT-5.4Turns=1, Tool Setup=Web Search2026.07 | 72.5 | 79 | 81 | 44 | 60.2 | 33.5 | 35.4 | 2 | 1 | 0.3703 | |
| o4-mini DeepResearchTurns=1*, Tool Setup=Web Search2026.07 | 72 | 81.5 | 84.5 | 46.3 | 53 | 29.2 | 31.7 | 0.1 | 0.47 | 6.0903 | |
| Qwen3.5-397BTurns=1, Tool Setup=Workflow2026.07 | 70 | 78 | 82 | 41.1 | 53.6 | 28.4 | 29.7 | 30 | 1.97 | 0.0233 | |
| Qwen3.5-9BTurns=1, Tool Setup=Toolset2026.07 | 69 | 82 | 85.5 | 38.3 | 56.5 | 31.2 | 33.4 | 168.3 | 3.03 | 0.0041 | |
| Qwen3.5-9BTurns=M, Tool Setup=Toolset2026.07 | 58 | 71 | 75.5 | 34.8 | 47.5 | 26.8 | 29 | 43.3 | 1.67 | 0.0134 | |
| GPT-5.4Turns=M, Tool Setup=Workflow2026.07 | 56.5 | 78 | 84 | 53.6 | 38.2 | 20.5 | 24 | 2.5 | 0.82 | 0.2231 | |
| GPT-5.4Turns=1, Tool Setup=Workflow2026.07 | 52 | 70 | 83 | 54.1 | 36.4 | 18.8 | 22.1 | 2.9 | 1.1 | 0.1803 | |
| Qwen3.5-9BTurns=1, Tool Setup=Workflow2026.07 | 48.5 | 66.5 | 77 | 53.7 | 31.2 | 17.4 | 21.5 | 147 | 2.93 | 0.0033 | |
| Qwen3.5-9BTurns=M, Tool Setup=Workflow2026.07 | 48 | 67.5 | 75 | 52.6 | 32.7 | 17.4 | 20.4 | 88.9 | 2 | 0.0054 |