Llama3_ko_4.2_sft: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1418 ± 17, rank #1768 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-CKA | 30.98 | Normalized accuracy (%) | 85.7 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.76 | Normalized accuracy (%) | 81.3 |
| Open Korean LLM - Ko-EQ-Bench | 12.5 | EQ-Bench score | 58.3 |
| Open Korean LLM - Ko-Winogrande | 59.75 | Accuracy (%) | 48.8 |
| Open Korean LLM Leaderboard | 34.53 | Average Score (%) | 48.1 |
| Open Ko-LLM Leaderboard | 34.53 | Average (%) | 48 |
| Open Korean LLM - KorNAT-Harmless | 62.39 | Normalized accuracy (%) | 43.3 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 10.77 | Exact match, flexible extract (%) | 35.1 |
| Open Korean LLM - KorNAT-Helpful | 44.95 | Normalized accuracy (%) | 33.4 |
| Open Korean LLM - Ko-IFEval | 22.13 | Strict accuracy, prompt/instruction mean (%) | 22.7 |
Interactive version: theaggregate.ai/model?slug=llama3-ko-4-2-sft · How It Works · Data refreshed daily, snapshot 2026-09-19.