GPQA gpqa Leaderboard
Auto-discovered from papers reporting GPQA (Accuracy). Β· Metric: Accuracy (higher is better) Β· π’ Updated 2h ago
| # | Model | Accuracy | Paper |
|---|---|---|---|
| 1 | + GRPO-Ξ» | 56.80 | link |
| 2 | +GRPO | 55.80 | link |
| 3 | Vanilla | 55.60 | link |
| 4 | +LP | 55.40 | link |
| 5 | RL with Learnable Textual Feedback: A Bilevel Approach | 49.30 | link |