Llama 2 13B .kor.v2: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1380 ± 17, rank #2052 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-Winogrande | 63.61 | Accuracy (%) | 75.1 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 13.87 | Exact match, flexible extract (%) | 42 |
| Open Korean LLM - KorNAT-Helpful | 44.82 | Normalized accuracy (%) | 32.9 |
| Open Korean LLM Leaderboard | 32.28 | Average Score (%) | 32.4 |
| Open Ko-LLM Leaderboard | 32.28 | Average (%) | 32.2 |
| Open Korean LLM - Ko-IFEval | 24.65 | Strict accuracy, prompt/instruction mean (%) | 30.9 |
| Open Korean LLM - Ko-EQ-Bench | 0 | EQ-Bench score | 24.5 |
| Open Korean LLM - KorNAT-Harmless | 60.97 | Normalized accuracy (%) | 21.5 |
| Open Korean LLM - KorNAT-CKA | 17.79 | Normalized accuracy (%) | 16.7 |
| Open Korean LLM - Ko-GPQA-Diamond | 19.19 | Normalized accuracy (%) | 6.4 |
Interactive version: theaggregate.ai/model?slug=llama-2-13b-kor-v2 · How It Works · Data refreshed daily, snapshot 2026-09-19.