llama-3-Korean-8B-V2: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1486 ± 17, rank #1214 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 46.41 | Strict accuracy, prompt/instruction mean (%) | 91.2 |
| Open Korean LLM - KorNAT-Helpful | 49.88 | Normalized accuracy (%) | 82.3 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.76 | Normalized accuracy (%) | 81.3 |
| Open Korean LLM Leaderboard | 40.88 | Average Score (%) | 79.2 |
| Open Ko-LLM Leaderboard | 40.88 | Average (%) | 79 |
| Open Korean LLM - KorNAT-Harmless | 64.31 | Normalized accuracy (%) | 72.7 |
| Open Korean LLM - Ko-EQ-Bench | 25.42 | EQ-Bench score | 71.4 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 25.4 | Exact match, flexible extract (%) | 58.9 |
| Open Korean LLM - KorNAT-CKA | 24.32 | Normalized accuracy (%) | 51.1 |
| Open Korean LLM - Ko-Winogrande | 59.19 | Accuracy (%) | 44.4 |
Interactive version: theaggregate.ai/model?slug=llama-3-korean-8b-v2 · How It Works · Data refreshed daily, snapshot 2026-09-19.