Llama-3-instruction-constructionsafety: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1452 ± 17, rank #1511 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-CKA | 36.82 | Normalized accuracy (%) | 97 |
| Open Korean LLM - Ko-Winogrande | 66.77 | Accuracy (%) | 89.5 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Ko-LLM Leaderboard | 37.9 | Average (%) | 64.2 |
| Open Korean LLM Leaderboard | 37.9 | Average Score (%) | 64.2 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 24.64 | Exact match, flexible extract (%) | 57.8 |
| Open Korean LLM - Ko-EQ-Bench | 7.69 | EQ-Bench score | 54 |
| Open Korean LLM - Ko-IFEval | 28.06 | Strict accuracy, prompt/instruction mean (%) | 46.2 |
| Open Korean LLM - KorNAT-Harmless | 61.1 | Normalized accuracy (%) | 23 |
| Open Korean LLM - KorNAT-Helpful | 42.45 | Normalized accuracy (%) | 16.6 |
Interactive version: theaggregate.ai/model?slug=llama-3-instruction-constructionsafety · How It Works · Data refreshed daily, snapshot 2026-09-19.