Scientific Literature Search Human Study
74.7Success Rate (SR)PAPERPILOT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PAPERPILOTworkflow type=multi-turn literature search agent2026.07 | 74.7 | 2.4 | 4.2 | 3.8 | |
| GPT-5.42026.07 | 32 | 7.8 | 2.4 | 4 | |
| OpenAI-DeepResearchevaluation protocol=one-shot system2026.07 | 8 | 27.4 | — | 1 |