MERA Original Paper - MathLogicQA: leaderboard
Metric: Accuracy (%; 5-shot, log-likelihood). Source: arxiv.org. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Yi 6B (Base) | 38.2 |
| 2 | davinci-002 | 35.3 |
| 3 | Mistral-7B-v0.1 | 34.4 |
| 4 | Llama 2 13B Base | 31.4 |
| 5 | Llama 2 7B Base | 27.7 |
| 6 | mGPT-13B | 26.3 |
| 7 | ruT5-base | 25.9 |
| 8 | mGPT | 25.8 |
| 9 | ruGPT-3.5 | 25.8 |
| 10 | ruGPT3-large | 25.1 |
| 11 | ruGPT-3 (Medium) | 24.8 |
| 12 | ruGPT-3-small | 24.4 |
Interactive version: theaggregate.ai/benchmark?slug=mera-original-paper-mathlogicqa · How It Works · Data refreshed daily, snapshot 2026-09-25.