Llama-2-kor-13B: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1406 ± 17, rank #1864 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-Winogrande | 62.59 | Accuracy (%) | 69.2 |
| Open Korean LLM - Ko-EQ-Bench | 13.76 | EQ-Bench score | 59.5 |
| Open Korean LLM - KorNAT-Harmless | 62.71 | Normalized accuracy (%) | 49.2 |
| Open Korean LLM - KorNAT-Helpful | 46.4 | Normalized accuracy (%) | 45.7 |
| Open Korean LLM Leaderboard | 33.71 | Average Score (%) | 43.7 |
| Open Ko-LLM Leaderboard | 33.71 | Average (%) | 43.5 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 11.22 | Exact match, flexible extract (%) | 36 |
| Open Korean LLM - Ko-IFEval | 25.08 | Strict accuracy, prompt/instruction mean (%) | 33 |
| Open Korean LLM - KorNAT-CKA | 21.12 | Normalized accuracy (%) | 32.9 |
| Open Korean LLM - Ko-GPQA-Diamond | 20.71 | Normalized accuracy (%) | 18.8 |
Interactive version: theaggregate.ai/model?slug=llama-2-kor-13b · How It Works · Data refreshed daily, snapshot 2026-09-19.