Rulebook Revision on Board Game Rulebook 207 games (test)
95.9Improvement %BG-Realizer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BG-RealizerMode=Single-pass2026.06 | 95.9 | 32.9 | |
| GPT-5.4Mode=Single-pass2026.06 | 93.6 | 13.1 | |
| Gemini-3.1-FlashMode=Single-pass2026.06 | 91.4 | 17.5 | |
| Qwen3.5-397BMode=Single-pass2026.06 | 91.2 | 12.6 | |
| Qwen3.5-27BMode=Single-pass2026.06 | 90 | 17.4 |