Large-scale QA benchmark with trivia questions and independently gathered evidence documents.
Accuracy is the reported evaluation metric for TriviaQA. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.
Higher is better
Muted rows were not state of the art when published — an earlier or same-year result already scored better.
| Rank | Model | Trust | Score | Year | Links | Fix |
|---|---|---|---|---|---|---|
| 01 | Llama 2 70B (5-shot) | unverified | 85 | 2023 | Paper ↗Code ↗ | Looks wrong? |
| 02 | LLaMA-65B | unverified | 73 | 2023 | Paper ↗Code ↗ | Looks wrong? |
| 03 | SmoLM2 (1.7B) | unverified | 36.7 | 2025 | Paper ↗Code ↗ | Looks wrong? |
| 04 | BitNet b1.58 2B4T | unverified | 33.57 | 2025 | Paper ↗Code ↗ | Looks wrong? |