Reasoning over Large Structured Context on RB-JSON
4.92ReasoningJudge ScoreGPT-5 + HYVE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5 + HYVEOptimization=HYVE pipeline2026.04 | 4.92 | 132.7 | 7.19 | |
| GPT-4.1 + HYVEOptimization=HYVE pipeline2026.04 | 4.92 | 107.3 | 2.11 | |
| GPT-5Optimization=Standard JSON serialization2026.04 | 4.84 | 142.8 | 7.39 | |
| GPT-4.1Optimization=Standard JSON serialization2026.04 | 4.78 | 116.1 | 2.4 |