pythia-12B-deduped: benchmark results
Provider: EleutherAI. Access: Open.
Aggregate Rating 1144 ± 16, rank #10045 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - High School Mathematics | 27.41 | Accuracy (%) | 54.7 |
| MMLU-by-task - Abstract Algebra | 29 | Accuracy (%) | 45.4 |
| MMLU-by-task - International Law | 46.28 | Accuracy (%) | 36.8 |
| MMLU-by-task - HellaSwag | 51.83 | Accuracy (%) | 34 |
| MMLU-by-task - World Religions | 35.09 | Accuracy (%) | 33.6 |
| MMLU-by-task - Professional Psychology | 29.25 | Accuracy (%) | 33.4 |
| MMLU-by-task - Moral Disputes | 30.92 | Accuracy (%) | 33.3 |
| MMLU-by-task - Econometrics | 26.32 | Accuracy (%) | 32.9 |
| MMLU-by-task - Elementary Mathematics | 26.46 | Accuracy (%) | 32.7 |
| MMLU-by-task - Moral Scenarios | 24.36 | Accuracy (%) | 32.6 |
| MMLU-by-task - Human Aging | 36.77 | Accuracy (%) | 32.1 |
| MMLU-by-task - Nutrition | 29.74 | Accuracy (%) | 31 |
Interactive version: theaggregate.ai/model?slug=pythia-12b-deduped · How It Works · Data refreshed daily, snapshot 2026-11-08.