Generation on VPTT-Bench 1.0 (test)
0.644VPTT Score (Novelty Adjusted)Comb. (BRAG + VPRAG)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Comb. (BRAG + VPRAG)Model=4o-mini2026.01 | 0.644 | — | |
| VPRAGModel=4o-mini2026.01 | 0.64 | 0.1 | |
| Comb. (BRAG + VPRAG)Model=Gemini2026.01 | 0.632 | — | |
| VPRAGModel=Qwen2026.01 | 0.631 | — | |
| BRAGModel=4o-mini2026.01 | 0.628 | 0.5 | |
| VPRAGModel=Gemini2026.01 | 0.625 | 0.2 | |
| BRAGModel=Gemini2026.01 | 0.616 | 0.3 | |
| Comb. (BRAG + VPRAG)Model=Qwen2026.01 | 0.602 | 0.7 | |
| BRAGModel=Qwen2026.01 | 0.581 | 1.1 | |
| Persona OnlyModel=4o-mini2026.01 | 0.402 | 8.4 | |
| Persona OnlyModel=Qwen2026.01 | 0.389 | 8.3 | |
| Persona OnlyModel=Gemini2026.01 | 0.379 | 7.1 | |
| BaselineModel=Qwen2026.01 | 0.316 | 11.9 | |
| BaselineModel=4o-mini2026.01 | 0.316 | 12.6 | |
| BaselineModel=Gemini2026.01 | 0.316 | 9.8 |