Mistral-7B-Base-SFT-RRHF: benchmark results
Princeton NLP's 7B Mistral checkpoint tuned with supervised fine-tuning and RRHF preference optimization, released alongside their SimPO preference-optimization research. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1353 ± 19, rank #1379 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GPQA-Diamond | 25.25 | Normalized accuracy (%) | 75.9 |
| Open Korean LLM - Ko-IFEval | 34.17 | Strict accuracy, prompt/instruction mean (%) | 70 |
| Open LLM Leaderboard - MuSR | 41.87 | Score | 59.2 |
| Open Korean LLM - Ko-EQ-Bench | 8.8 | EQ-Bench score | 54.9 |
| Open LLM Leaderboard - IFEval | 44.07 | Score | 47.5 |
| Open LLM Leaderboard - GPQA | 29.03 | Score | 45 |
| Open Korean LLM Leaderboard | 33.6 | Average Score (%) | 43.1 |
| Open Ko-LLM Leaderboard | 33.6 | Average (%) | 42.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 14.4 | Exact match, flexible extract (%) | 42.8 |
| Open Korean LLM - KorNAT-Helpful | 45.68 | Normalized accuracy (%) | 38.5 |
| Open Portuguese LLM - ENEM | 51.08 | Accuracy (%) | 35 |
| Open Portuguese LLM - ASSIN2 RTE | 84.17 | Macro F1 (%) | 34 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-base-sft-rrhf · How It Works · Data refreshed daily, snapshot 2026-10-23.