alpaca-native: benchmark results
maicomputer's alpaca-native, a replica of Stanford's original Alpaca model trained with a minor FSDP-mode modification to the original recipe. Provider: maicomputer. Released 2023-03-16. Access: Open.
Aggregate Rating 1242 ± 15, rank #9317 of 10916 rated models, from 78 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MMLU-by-task - College Mathematics | 35 | Accuracy (%) | 75 |
| MMLU-by-task - Machine Learning | 36.61 | Accuracy (%) | 68 |
| MMLU-by-task - College Computer Science | 41 | Accuracy (%) | 57.6 |
| MMLU-by-task - High School Physics | 31.13 | Accuracy (%) | 56.1 |
| MMLU-by-task - ARC Challenge | 51.28 | Accuracy (%) | 55.9 |
| MMLU-by-task - HellaSwag | 59.6 | Accuracy (%) | 55.1 |
| LLM Trustworthy - Adversarial | 46.43 | Trust Score (%) | 54.2 |
| MMLU-by-task - Anatomy | 45.93 | Accuracy (%) | 53.8 |
| MMLU-by-task - High School Computer Science | 45 | Accuracy (%) | 52.2 |
| MMLU-by-task - Global Facts | 33 | Accuracy (%) | 51.6 |
| MMLU-by-task - Elementary Mathematics | 28.84 | Accuracy (%) | 50.7 |
| LLM Trustworthy - Fairness | 92.63 | Trust Score (%) | 50 |
Interactive version: theaggregate.ai/model?slug=alpaca-native · How It Works · Data refreshed daily, snapshot 2026-11-08.