Llama-3-Instruct-8B-ORPO-v0.2: benchmark results
Princeton NLP's ORPO preference-tuned Llama 3 8B Instruct, released as a baseline from the SimPO paper. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1462 ± 19, rank #927 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-EQ-Bench | 41.43 | EQ-Bench score | 92 |
| Open LLM Leaderboard - IFEval | 76.33 | Score | 92 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 55.04 | Exact match, flexible extract (%) | 88.9 |
| Open Korean LLM Leaderboard | 42.97 | Average Score (%) | 87.1 |
| Open Ko-LLM Leaderboard | 42.97 | Average (%) | 87 |
| Open Korean LLM - Ko-IFEval | 41.26 | Strict accuracy, prompt/instruction mean (%) | 84.1 |
| Open Portuguese LLM - OAB Exams | 49.79 | Accuracy (%) | 74.8 |
| Open Portuguese LLM - ENEM | 69.63 | Accuracy (%) | 73.5 |
| Open Korean LLM - KorNAT-Helpful | 48.83 | Normalized accuracy (%) | 73.2 |
| Open Portuguese LLM - BLUEX | 58.28 | Accuracy (%) | 72.6 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Portuguese LLM - ASSIN2 RTE | 91.61 | Macro F1 (%) | 68.7 |
Interactive version: theaggregate.ai/model?slug=llama-3-instruct-8b-orpo-v0-2 · How It Works · Data refreshed daily, snapshot 2026-10-23.