LLaMA-65B (Instruct): benchmark results
Provider: Meta. Access: Open.
Aggregate Rating 1404 ± 16, rank #5908 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - College Physics | 45.1 | Accuracy (%) | 99.8 |
| MMLU-by-task - Public Relations | 73.64 | Accuracy (%) | 97.6 |
| MMLU-by-task - TruthfulQA MC1 | 42.96 | Accuracy (%) | 97.4 |
| MMLU-by-task - TruthfulQA MC2 | 59.7 | Accuracy (%) | 97.2 |
| MMLU-by-task - ARC Challenge | 65.27 | Accuracy (%) | 96.4 |
| MMLU-by-task - Management | 82.52 | Accuracy (%) | 96.4 |
| MMLU-by-task - Virology | 53.61 | Accuracy (%) | 95.4 |
| MMLU-by-task - Security Studies | 77.96 | Accuracy (%) | 95.1 |
| MMLU-by-task - Moral Scenarios | 49.05 | Accuracy (%) | 94.4 |
| MMLU-by-task - Moral Disputes | 75.14 | Accuracy (%) | 94.1 |
| MMLU-by-task - Professional Accounting | 52.84 | Accuracy (%) | 94.1 |
| MMLU-by-task - Clinical Knowledge | 69.81 | Accuracy (%) | 93.7 |
Interactive version: theaggregate.ai/model?slug=llama-65b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.