Mistral-7B-Base-SFT-CPO: benchmark results
Princeton NLP's 7B Mistral checkpoint tuned with supervised fine-tuning and Contrastive Preference Optimization, released alongside their SimPO preference-optimization research. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1376 ± 20, rank #1311 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-EQ-Bench | 26.57 | EQ-Bench score | 73.4 |
| Open Korean LLM - Ko-IFEval | 35.12 | Strict accuracy, prompt/instruction mean (%) | 72.3 |
| Open Portuguese LLM - FaQuAD NLI | 73.34 | Macro F1 (%) | 68.4 |
| Open Korean LLM Leaderboard | 35.92 | Average Score (%) | 54.8 |
| Open Ko-LLM Leaderboard | 35.92 | Average (%) | 54.7 |
| Open Korean LLM - KorNAT-Harmless | 63.01 | Normalized accuracy (%) | 53.5 |
| Open LLM Leaderboard - IFEval | 46.55 | Score | 52.9 |
| Open LLM Leaderboard - MuSR | 40.71 | Score | 47.8 |
| Open LLM Leaderboard - GPQA | 29.19 | Score | 46.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 15.92 | Exact match, flexible extract (%) | 46.1 |
| Open Portuguese LLM - BLUEX | 48.12 | Accuracy (%) | 41.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 22.73 | Normalized accuracy (%) | 40.9 |
Interactive version: theaggregate.ai/model?slug=mistral-7b-base-sft-cpo · How It Works · Data refreshed daily, snapshot 2026-10-23.