LLaMA3-iterative-DPO-final: benchmark results
Provider: Meta. Released 2024-05-17. Access: Open.
Unified ELO 1405 ± 1, rank #1116 of 1392 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM Leaderboard | 46.05 | Average Score (%) | 92.5 |
| Open LLM Leaderboard - IFEval | 53.4 | Score | 63.7 |
| Open LLM Leaderboard - BBH | 29.79 | Score | 50.9 |
| RewardBench | 67.83 | Score (%) | 50.5 |
| Open LLM Leaderboard - MMLU-Pro | 25.08 | Score | 44.9 |
| RewardBench Chat Hard | 59.21 | Accuracy (%) | 44.3 |
| Open LLM Leaderboard - MATH Level 5 | 8.84 | Score | 43.6 |
| RewardBench Safety | 78.65 | Accuracy (%) | 41.5 |
| Open LLM Leaderboard - GPQA | 4.47 | Score | 38.5 |
| Open LLM Leaderboard - MuSR | 5.08 | Score | 24.9 |
| RewardBench Chat | 83.8 | Accuracy (%) | 17.6 |
| RewardBench Reasoning | 61.61 | Accuracy (%) | 17.2 |
Interactive version: theaggregate.ai/model?slug=llama3-iterative-dpo-final · How It Works · Data refreshed daily, snapshot 2026-09-05.