Interactive Question Answering on QuestBench Math
53.9AccuracyOursI
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OursIEvaluation Protocol=Multi-turn2026.02 | 53.9 | 83.5 | 9.7 | |
| OursOEvaluation Protocol=Multi-turn2026.02 | 38.8 | 68.2 | 35.4 | |
| GeminiEvaluation Protocol=Multi-turn2026.02 | 35.4 | 33.5 | 4.4 | |
| GPTEvaluation Protocol=Multi-turn2026.02 | 32 | 31.6 | 2.4 | |
| QwenEvaluation Protocol=Multi-turn2026.02 | 32 | 37.9 | 0.5 | |
| FATAEvaluation Protocol=Multi-turn2026.02 | 32 | 41.9 | 1.5 | |
| AskToActEvaluation Protocol=Multi-turn2026.02 | 18 | 27.2 | 5.3 |