Vicuna-13B-CoT-fp16: benchmark results
TheBloke's Hugging Face publication of Kevin Pro's Vicuna 13B CoT, whose original release is kevinpro/Vicuna-13B-CoT. Conversion publication is credited separately from the original model creator. Provider: TheBloke. Access: Open.
Aggregate Rating 1316 ± 15, rank #8414 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - Formal Logic | 40.48 | Accuracy (%) | 87.2 |
| MMLU-by-task - Human Sexuality | 67.18 | Accuracy (%) | 86.1 |
| MMLU-by-task - TruthfulQA MC1 | 36.23 | Accuracy (%) | 86.1 |
| MMLU-by-task - TruthfulQA MC2 | 52.08 | Accuracy (%) | 85.4 |
| MMLU-by-task - Sociology | 77.11 | Accuracy (%) | 85.2 |
| MMLU-by-task - Machine Learning | 39.29 | Accuracy (%) | 79 |
| MMLU-by-task - Elementary Mathematics | 33.86 | Accuracy (%) | 78.1 |
| MMLU-by-task - College Biology | 59.72 | Accuracy (%) | 77.1 |
| MMLU-by-task - High School Mathematics | 30 | Accuracy (%) | 76.1 |
| MMLU-by-task - College Computer Science | 46 | Accuracy (%) | 75.6 |
| MMLU-by-task - Medical Genetics | 59 | Accuracy (%) | 75.4 |
| MMLU-by-task - Security Studies | 62.45 | Accuracy (%) | 74.9 |
Interactive version: theaggregate.ai/model?slug=vicuna-13b-cot-fp16 · How It Works · Data refreshed daily, snapshot 2026-11-08.