Mistral-7B-Base-SFT-RDPO: benchmark results
Princeton NLP's 7B Mistral checkpoint fine-tuned with supervised fine-tuning and RDPO preference optimization, released alongside their SimPO preference-optimization research. Provider: Princeton NLP. Released 2024-05-17. Access: Open.
Aggregate Rating 1367 ± 19, rank #1347 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GPQA-Diamond | 28.79 | Normalized accuracy (%) | 98.1 |
| Open Korean LLM - Ko-IFEval | 37.05 | Strict accuracy, prompt/instruction mean (%) | 77.4 |
| Open Korean LLM - KorNAT-Helpful | 48.58 | Normalized accuracy (%) | 69.9 |
| Open Korean LLM - Ko-EQ-Bench | 17.08 | EQ-Bench score | 62.8 |
| Open LLM Leaderboard - IFEval | 46.06 | Score | 51.9 |
| Open Portuguese LLM - FaQuAD NLI | 66.15 | Macro F1 (%) | 51.8 |
| Open Korean LLM Leaderboard | 34.55 | Average Score (%) | 48.4 |
| Open Ko-LLM Leaderboard | 34.55 | Average (%) | 48.2 |
| Open Portuguese LLM - ASSIN2 RTE | 89.08 | Macro F1 (%) | 47.9 |
| Open Korean LLM - KorNAT-Harmless | 62.62 | Normalized accuracy (%) | 47.4 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 15.09 | Exact match, flexible extract (%) | 44.4 |
| Open Portuguese LLM - OAB Exams | 40.14 | Accuracy (%) | 43.2 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-base-sft-rdpo · How It Works · Data refreshed daily, snapshot 2026-10-23.