Llama 2 13B Instruct: benchmark results
Provider: Meta. Released 2023-07-18. Access: Open.
Aggregate Rating 1345 ± 15, rank #7836 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - HellaSwag | 65.19 | Accuracy (%) | 89.6 |
| MMLU-by-task - Professional Accounting | 47.52 | Accuracy (%) | 89.4 |
| MMLU-by-task - Clinical Knowledge | 63.02 | Accuracy (%) | 88.3 |
| MMLU-by-task - Prehistory | 68.83 | Accuracy (%) | 87.9 |
| MMLU-by-task - High School US History | 78.43 | Accuracy (%) | 84.7 |
| MMLU-by-task - Professional Psychology | 58.82 | Accuracy (%) | 84.2 |
| MMLU-by-task - College Biology | 62.5 | Accuracy (%) | 84 |
| MMLU-by-task - Elementary Mathematics | 34.92 | Accuracy (%) | 82.7 |
| MMLU-by-task - Medical Genetics | 62 | Accuracy (%) | 82.4 |
| MMLU-by-task - Professional Law | 43.48 | Accuracy (%) | 82.4 |
| MMLU-by-task - High School Chemistry | 44.33 | Accuracy (%) | 82.2 |
| MMLU-by-task - Moral Scenarios | 40.67 | Accuracy (%) | 82.2 |
Interactive version: theaggregate.ai/model?slug=llama-2-13b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.