Llama-3.1-Korean-8B-Instruct: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1508 ± 17, rank #1043 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 45 | Strict accuracy, prompt/instruction mean (%) | 89.4 |
| Open Korean LLM Leaderboard | 42.43 | Average Score (%) | 86.5 |
| Open Ko-LLM Leaderboard | 42.43 | Average (%) | 86.4 |
| Open Korean LLM - Ko-GPQA-Diamond | 26.26 | Normalized accuracy (%) | 85.5 |
| Open Korean LLM - Ko-EQ-Bench | 30.67 | EQ-Bench score | 79.7 |
| Open Korean LLM - KorNAT-CKA | 29.04 | Normalized accuracy (%) | 78.5 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 36.01 | Exact match, flexible extract (%) | 69.2 |
| Open Korean LLM - Ko-Winogrande | 59.59 | Accuracy (%) | 47.5 |
| Open Korean LLM - KorNAT-Helpful | 45.48 | Normalized accuracy (%) | 37.1 |
| Open Korean LLM - KorNAT-Harmless | 60.77 | Normalized accuracy (%) | 19 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-korean-8b-instruct · How It Works · Data refreshed daily, snapshot 2026-09-19.