Llama 2 70B Instruct: benchmark results
Provider: Meta. Released 2023-07-18. Access: Open.
Aggregate Rating 1439 ± 17, rank #4215 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - Abstract Algebra | 41 | Accuracy (%) | 99.5 |
| MMLU-by-task - Clinical Knowledge | 75.47 | Accuracy (%) | 99.5 |
| MMLU-by-task - College Biology | 84.72 | Accuracy (%) | 99.5 |
| MMLU-by-task - College Physics | 43.14 | Accuracy (%) | 99.3 |
| MMLU-by-task - Miscellaneous | 87.36 | Accuracy (%) | 99.3 |
| MMLU-by-task - World Religions | 87.72 | Accuracy (%) | 99.3 |
| MMLU-by-task - Business Ethics | 76 | Accuracy (%) | 99.1 |
| MMLU-by-task - Prehistory | 83.95 | Accuracy (%) | 99 |
| MMLU-by-task - Professional Psychology | 76.31 | Accuracy (%) | 98.9 |
| MMLU-by-task - High School Geography | 88.89 | Accuracy (%) | 98.8 |
| MMLU-by-task - Conceptual Physics | 68.94 | Accuracy (%) | 98.7 |
| MMLU-by-task - College Mathematics | 42 | Accuracy (%) | 98.4 |
Interactive version: theaggregate.ai/model?slug=llama-2-70b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.