Llama 2 13B FP16: benchmark results
A community fp16 conversion by TheBloke of Meta's original Llama 2 13B base model, converted from Meta's PTH weights to Hugging Face Transformers format. Provider: TheBloke. Released 2023-07-18. Access: Open.
Aggregate Rating 1340 ± 14, rank #7963 of 10916 rated models, from 77 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - High School Physics | 37.09 | Accuracy (%) | 88.1 |
| MMLU-by-task - Abstract Algebra | 35 | Accuracy (%) | 87.4 |
| MMLU-by-task - College Chemistry | 44 | Accuracy (%) | 86.5 |
| MMLU-by-task - Jurisprudence | 74.07 | Accuracy (%) | 85.9 |
| MMLU-by-task - High School Chemistry | 44.83 | Accuracy (%) | 84.1 |
| MMLU-by-task - Moral Disputes | 64.74 | Accuracy (%) | 83.8 |
| MMLU-by-task - High School Biology | 67.42 | Accuracy (%) | 83 |
| MMLU-by-task - Nutrition | 62.42 | Accuracy (%) | 82.8 |
| MMLU-by-task - Computer Security | 70 | Accuracy (%) | 82.7 |
| MMLU-by-task - Formal Logic | 38.1 | Accuracy (%) | 82 |
| MMLU-by-task - High School Microeconomics | 58.4 | Accuracy (%) | 82 |
| MMLU-by-task - College Medicine | 53.18 | Accuracy (%) | 81.9 |
Interactive version: theaggregate.ai/model?slug=llama-2-13b-fp16 · How It Works · Data refreshed daily, snapshot 2026-11-08.