Retrieval-Augmented Question Answering on 170 unique canonical user-facing queries Complex Rec. n=30
72FaithfulnessI-GUIDE SS
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| I-GUIDE SSmode=Smart Search2026.06 | 72 | 63 | 64 | 64 | |
| Claude Opus 4.72026.06 | 68 | — | 72 | 69 | |
| GPT-5.52026.06 | 67 | — | 66 | 70 | |
| Naive RAG2026.06 | 66 | 23 | 30 | 31 | |
| Qwen2.52026.06 | 58 | — | 34 | 50 | |
| GPT-4o2026.06 | 52 | — | 47 | 52 |