Large-scale trivia question answering — 650K question-evidence-answer triples from trivia enthusiasts, requiring both reading comprehension and world knowledge.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Large, diverse question set. Tests both knowledge recall and evidence-based reasoning.
Non-standard score range in Epoch data (1–64). Older benchmark likely in training data.
Higher is better. Score interpretation varies by evaluation method.
Non-standard score range in Epoch data (1–64).
| # | Model | Score |
|---|---|---|
| 1 | 80.3 | |
| 2 | 64.1 | |
| 3 | 56.4 | |
| 4 | 48.1 | |
| 5 | 37.0 | |
| 6 | 31.9 | |
| 7 | 31.6 | |
| 8 | 31.4 | |
| 9 | 29.8 | |
| 10 | 29.7 | |
| 11 | 29.3 | |
| 12 | 28.2 | |
| 13 | 27.7 | |
| 14 | 27.6 | |
| 15 | 26.3 | |
| 16 | 25.8 | |
| 17 | 25.6 | |
| 18 | 25.5 | |
| 19 | 24.3 | |
| 20 | 24.3 | |
| 21 | 24.3 | |
| 22 | 21.9 | |
| 23 | 21.6 | |
| 24 | 21.4 | |
| 25 | 21.3 | |
| 26 | 21.1 | |
| 27 | 20.8 | |
| 28 | 20.5 | |
| 29 | 20.4 | |
| 30 | 20.3 | |
| 31 | 19.8 | |
| 32 | 19.7 | |
| 33 | 19.7 | |
| 34 | 19.5 | |
| 35 | 19.4 | |
| 36 | 19.4 | |
| 37 | 19.4 | |
| 38 | 18.9 | |
| 39 | 18.3 | |
| 40 | 18.1 | |
| 41 | 18.1 | |
| 42 | 18.1 | |
| 43 | 18.1 | |
| 44 | 18.0 | |
| 45 | 17.9 | |
| 46 | 17.8 | |
| 47 | 17.5 | |
| 48 | 17.4 | |
| 49 | 17.1 | |
| 50 | 16.8 |