Open-Domain QA on AmbigQA Nq=300
0.473AccGemini3Pro
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Gemini3ProNq=3002026.02 | 0.473 | 0.538 | 0.48 | 0.659 | 0.738 | 0.326 | 0.412 | |
| KimiK2Nq=3002026.02 | 0.407 | 0.663 | 0.466 | 0.77 | 0.761 | 0.163 | 0.597 | |
| Claude4.5SNq=3002026.02 | 0.357 | 0.501 | 0.475 | 0.533 | 0.407 | 0.357 | 0.05 | |
| PooledNq=15002026.02 | 0.351 | 0.55 | 0.498 | 0.627 | 0.566 | 0.211 | 0.355 | |
| GPT-OSSNq=3002026.02 | 0.27 | 0.641 | 0.524 | 0.702 | 0.538 | 0.081 | 0.457 | |
| Qwen3Nq=3002026.02 | 0.25 | 0.476 | 0.55 | 0.506 | 0.298 | 0.237 | 0.062 |