Mistral-7B-Instruct-ORPO: benchmark results
Princeton NLP's 7B Mistral Instruct checkpoint tuned with ORPO preference optimization, released as part of their SimPO preference-optimization research line. Provider: Princeton NLP. Released 2024-05-17. Access: Open.
Aggregate Rating 1371 ± 21, rank #1325 of 1654 rated models, from 20 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-Harmless | 66.53 | Normalized accuracy (%) | 86.2 |
| Open Korean LLM - KorNAT-Helpful | 50.35 | Normalized accuracy (%) | 85.4 |
| Open Korean LLM - Ko-EQ-Bench | 26.79 | EQ-Bench score | 73.6 |
| Open Korean LLM - Ko-IFEval | 32.65 | Strict accuracy, prompt/instruction mean (%) | 63.4 |
| Open Portuguese LLM - ASSIN2 RTE | 90.28 | Macro F1 (%) | 55.8 |
| Open Korean LLM Leaderboard | 35.95 | Average Score (%) | 54.6 |
| Open LLM Leaderboard - IFEval | 47.2 | Score | 54.2 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 20.32 | Exact match, flexible extract (%) | 52.1 |
| Open Portuguese LLM - FaQuAD NLI | 66.27 | Macro F1 (%) | 52.1 |
| Open Portuguese LLM - BLUEX | 50.49 | Accuracy (%) | 47.4 |
| Open Portuguese LLM - ENEM | 58.64 | Accuracy (%) | 44.2 |
| Open Portuguese LLM - OAB Exams | 39.23 | Accuracy (%) | 40.5 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-instruct-orpo · How It Works · Data refreshed daily, snapshot 2026-10-24.