Question Answering on TOON-QA
98Exact Match (EM)GPT-5 + HYVE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5 + HYVEOptimization=HYVE pipeline2026.04 | 98 | 182.4 | 2.71 | |
| GPT-5Optimization=Standard JSON serialization2026.04 | 96 | 1.4 | 5.48 | |
| GPT-4.1 + HYVEOptimization=HYVE pipeline2026.04 | 93 | 153.8 | 2.33 | |
| GPT-4.1Optimization=Standard JSON serialization2026.04 | 47 | 1.3 | 1.35 |