Llama-3-Instruct-8B-CPO-v0.2: benchmark results
Princeton NLP's Llama-3-Instruct-8B-CPO-v0.2, a Llama 3 8B Instruct model aligned via CPO, released alongside Princeton's SimPO preference-optimization paper. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1458 ± 19, rank #954 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-EQ-Bench | 41.78 | EQ-Bench score | 92.2 |
| Open Korean LLM - Ko-IFEval | 47.55 | Strict accuracy, prompt/instruction mean (%) | 91.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 56.63 | Exact match, flexible extract (%) | 90.4 |
| Open LLM Leaderboard - IFEval | 75.06 | Score | 90.1 |
| Open Ko-LLM Leaderboard | 42.21 | Average (%) | 85.2 |
| Open Korean LLM Leaderboard | 42.21 | Average Score (%) | 85.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.76 | Normalized accuracy (%) | 81.3 |
| Open Portuguese LLM - ENEM | 69.7 | Accuracy (%) | 73.7 |
| Open Korean LLM - KorNAT-Helpful | 48.85 | Normalized accuracy (%) | 73.6 |
| Open Portuguese LLM - OAB Exams | 49.61 | Accuracy (%) | 73.5 |
| Open Portuguese LLM - BLUEX | 58.28 | Accuracy (%) | 72.6 |
| Open Portuguese LLM - ASSIN2 RTE | 91.62 | Macro F1 (%) | 68.8 |
Interactive version: theaggregate.ai/model?slug=llama-3-instruct-8b-cpo-v0-2 · How It Works · Data refreshed daily, snapshot 2026-10-23.