sft_merged: benchmark results
Provider: Other. Access: Open.
Unified ELO 1391 ± 18, rank #1975 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-CKA | 32.46 | Normalized accuracy (%) | 88.9 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Korean LLM - Ko-Winogrande | 59.27 | Accuracy (%) | 45 |
| Open Korean LLM Leaderboard | 33.08 | Average Score (%) | 38.1 |
| Open Ko-LLM Leaderboard | 33.08 | Average (%) | 37.8 |
| Open Korean LLM - Ko-IFEval | 25.21 | Strict accuracy, prompt/instruction mean (%) | 33.6 |
| Open Korean LLM - KorNAT-Harmless | 61.45 | Normalized accuracy (%) | 28.9 |
| Open Korean LLM - Ko-EQ-Bench | 0 | EQ-Bench score | 24.5 |
| Open Korean LLM - KorNAT-Helpful | 42.73 | Normalized accuracy (%) | 18.3 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 1.59 | Exact match, flexible extract (%) | 14.5 |
Interactive version: theaggregate.ai/model?slug=sft-merged · How It Works · Data refreshed daily, snapshot 2026-09-19.