Deep Research on Natural Questions (NQ) (test)
97.8AccuracyOur best agent
Evaluation results
| Method | Links | |
|---|---|---|
| Our best agentType=Live web2025.10 | 97.8 | |
| WebSailorType=Live web2025.10 | 97.6 | |
| DeepResearcherType=Live web2025.10 | 89.6 | |
| Our base agentType=Live web2025.10 | 88.3 | |
| ASearcherType=Live web2025.10 | 87.2 | |
| Search-R1Type=Static & simulated web2025.10 | 79.2 | |
| R1-SearcherType=Static & simulated web2025.10 | 66.2 | |
| ZeroSearchType=Static & simulated web2025.10 | 30 |