llama-3-chat: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1440 ± 17, rank #1588 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GSM8K (flexible extract) | 49.43 | Exact match, flexible extract (%) | 82.9 |
| Open Korean LLM - Ko-IFEval | 39.62 | Strict accuracy, prompt/instruction mean (%) | 81.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Korean LLM - KorNAT-Harmless | 63.44 | Normalized accuracy (%) | 62.7 |
| Open Korean LLM - Ko-EQ-Bench | 11.46 | EQ-Bench score | 57.3 |
| Open Korean LLM - KorNAT-CKA | 25.27 | Normalized accuracy (%) | 56.5 |
| Open Korean LLM Leaderboard | 36.1 | Average Score (%) | 55.8 |
| Open Ko-LLM Leaderboard | 36.1 | Average (%) | 55.7 |
| Open Korean LLM - KorNAT-Helpful | 47.3 | Normalized accuracy (%) | 55.2 |
| Open Korean LLM - Ko-Winogrande | 60.22 | Accuracy (%) | 52.3 |
Interactive version: theaggregate.ai/model?slug=llama-3-chat · How It Works · Data refreshed daily, snapshot 2026-09-19.