Question Answering on FinanceBench Single-Document 9
84AccuracyDeepRead
Evaluation results
| Method | Links | |
|---|---|---|
| DeepReadSearch expansion=Enabled2026.02 | 84 | |
| Search-o1Search expansion=Enabled2026.02 | 83.3 | |
| DeepRead2026.02 | 82.7 | |
| Search-o12026.02 | 80 | |
| ITRGRetrieval refinement strategy=refine2026.02 | 52 | |
| ITRGRetrieval refinement strategy=refresh2026.02 | 48.1 | |
| Dense RAGReranking=Enabled2026.02 | 47.3 | |
| RAPTOR2026.02 | 38.7 | |
| Dense RAG2026.02 | 38.1 |