vicuna-13B-v1.5: benchmark results
LMSYS's 13B chat assistant, fine-tuned from Llama 2 on about 125,000 ShareGPT conversations for research use, released under the Llama 2 Community License. Provider: LMSYS. Released 2023-08-01. Access: Open.
Aggregate Rating 1346 ± 10, rank #1392 of 1656 rated models, from 112 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - Global Facts | 44 | Accuracy (%) | 93.2 |
| MMLU-by-task - College Medicine | 57.23 | Accuracy (%) | 90.3 |
| MMLU-by-task - Public Relations | 68.18 | Accuracy (%) | 88.8 |
| MMLU-by-task - Nutrition | 64.71 | Accuracy (%) | 88.3 |
| MMLU-by-task - Machine Learning | 44.64 | Accuracy (%) | 88.2 |
| MMLU-by-task - International Law | 76.86 | Accuracy (%) | 87.2 |
| MMLU-by-task - High School Mathematics | 31.85 | Accuracy (%) | 87 |
| MMLU-by-task - Clinical Knowledge | 62.26 | Accuracy (%) | 86.3 |
| MMLU-by-task - Sociology | 77.61 | Accuracy (%) | 86.1 |
| MMLU-by-task - Jurisprudence | 74.07 | Accuracy (%) | 85.9 |
| MMLU-by-task - Logical Fallacies | 69.33 | Accuracy (%) | 84.6 |
| MMLU-by-task - Marketing | 83.33 | Accuracy (%) | 84.3 |
Interactive version: theaggregate.ai/model?slug=vicuna-13b-v1-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.