llama2-chat-koalpaca: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1388 ± 17, rank #1993 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-EQ-Bench | 13.47 | EQ-Bench score | 59.2 |
| Open Korean LLM - Ko-IFEval | 29.97 | Strict accuracy, prompt/instruction mean (%) | 53.5 |
| Open Korean LLM - KorNAT-Harmless | 62.68 | Normalized accuracy (%) | 48.8 |
| Open Korean LLM - Ko-GPQA-Diamond | 22.73 | Normalized accuracy (%) | 40.9 |
| Open Korean LLM Leaderboard | 32.99 | Average Score (%) | 37.1 |
| Open Ko-LLM Leaderboard | 32.99 | Average (%) | 36.8 |
| Open Korean LLM - KorNAT-Helpful | 44.57 | Normalized accuracy (%) | 30.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 5.99 | Exact match, flexible extract (%) | 25.9 |
| Open Korean LLM - KorNAT-CKA | 18.01 | Normalized accuracy (%) | 17.5 |
| Open Korean LLM - Ko-Winogrande | 54.46 | Accuracy (%) | 16.3 |
Interactive version: theaggregate.ai/model?slug=llama2-chat-koalpaca · How It Works · Data refreshed daily, snapshot 2026-09-19.