Mistral-7B-Instruct-CPO: benchmark results
Princeton NLP's Mistral-7B-Instruct-CPO, a Mistral 7B Instruct model aligned via CPO, released alongside Princeton's SimPO preference-optimization paper. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1381 ± 19, rank #1291 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-Helpful | 50.78 | Normalized accuracy (%) | 87 |
| Open Korean LLM - KorNAT-Harmless | 66.41 | Normalized accuracy (%) | 85.8 |
| Open Korean LLM - Ko-EQ-Bench | 26.31 | EQ-Bench score | 73.1 |
| Open Portuguese LLM - FaQuAD NLI | 70.6 | Macro F1 (%) | 61.8 |
| Open Korean LLM Leaderboard | 37.17 | Average Score (%) | 61.1 |
| Open Ko-LLM Leaderboard | 37.17 | Average (%) | 61 |
| Open Portuguese LLM - ASSIN2 RTE | 90.56 | Macro F1 (%) | 58.7 |
| Open LLM Leaderboard - MuSR | 41.78 | Score | 58 |
| Open Korean LLM - Ko-IFEval | 30.36 | Strict accuracy, prompt/instruction mean (%) | 55.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 21.3 | Exact match, flexible extract (%) | 53.6 |
| Open Portuguese LLM - BLUEX | 51.04 | Accuracy (%) | 48.4 |
| Open Portuguese LLM - ENEM | 59.27 | Accuracy (%) | 45.2 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-instruct-cpo · How It Works · Data refreshed daily, snapshot 2026-10-23.