gemma-2-9B-it-SimPO: benchmark results
Princeton NLP's SimPO reference-free preference-optimization research tune of Gemma 2 9B IT on UltraFeedback-ArmoRM data. Provider: Princeton NLP. Released 2024-07-17. Access: Open.
Aggregate Rating 1561 ± 15, rank #580 of 1656 rated models, from 55 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GSM8K (flexible extract) | 72.02 | Exact match, flexible extract (%) | 98.5 |
| Open Korean LLM - Ko-IFEval | 65.34 | Strict accuracy, prompt/instruction mean (%) | 98.2 |
| Open Korean LLM - Ko-EQ-Bench | 58.77 | EQ-Bench score | 98.1 |
| Open Korean LLM - Ko-GPQA-Diamond | 28.79 | Normalized accuracy (%) | 98.1 |
| AlpacaEval 2.0 | 72.35 | LC Win Rate (%) | 96.8 |
| Open Ko-LLM Leaderboard | 49.27 | Average (%) | 96.5 |
| Open Korean LLM Leaderboard | 49.27 | Average Score (%) | 96.5 |
| Open Korean LLM - KorNAT-CKA | 36.32 | Normalized accuracy (%) | 96.4 |
| Open Portuguese LLM - ASSIN2 RTE | 93.82 | Macro F1 (%) | 92.2 |
| MEDIC Benchmark | 81.55 | MEDIC Public Table Average | 89.5 |
| Open CoT - LSAT Analytical Reasoning | 8.26 | CoT Gain (%) | 89.3 |
| Open Portuguese LLM - ENEM | 75.09 | Accuracy (%) | 88.6 |
Interactive version: theaggregate.ai/model?slug=gemma-2-9b-it-simpo · How It Works · Data refreshed daily, snapshot 2026-10-20.