llama3.1-korean-v1.4-sft-by-aidx: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1540 ± 17, rank #888 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM Leaderboard | 46.17 | Average Score (%) | 92.8 |
| Open Ko-LLM Leaderboard | 46.17 | Average (%) | 92.7 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 57.54 | Exact match, flexible extract (%) | 91.8 |
| Open Korean LLM - KorNAT-CKA | 33.29 | Normalized accuracy (%) | 91.5 |
| Open Korean LLM - Ko-EQ-Bench | 38.14 | EQ-Bench score | 89.1 |
| Open Korean LLM - Ko-IFEval | 40.55 | Strict accuracy, prompt/instruction mean (%) | 83.1 |
| Open Korean LLM - KorNAT-Harmless | 63.77 | Normalized accuracy (%) | 66.7 |
| Open Korean LLM - Ko-Winogrande | 60.93 | Accuracy (%) | 58.7 |
| Open Korean LLM - KorNAT-Helpful | 47.02 | Normalized accuracy (%) | 52.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 22.22 | Normalized accuracy (%) | 35 |
Interactive version: theaggregate.ai/model?slug=llama3-1-korean-v1-4-sft-by-aidx · How It Works · Data refreshed daily, snapshot 2026-09-19.