SOLAR-10.7B-Instruct-T3Q-ko-solar-dpo-ties: benchmark results
Provider: Upstage. Access: Open.
Unified ELO 1507 ± 18, rank #1054 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-Winogrande | 76.87 | Accuracy (%) | 99.9 |
| Open Korean LLM - KorNAT-Harmless | 88.04 | Normalized accuracy (%) | 99.9 |
| Open Korean LLM - KorNAT-Helpful | 79.47 | Normalized accuracy (%) | 99.8 |
| Open Korean LLM - KorNAT-CKA | 44.39 | Normalized accuracy (%) | 99.7 |
| Open Korean LLM - Ko-GPQA-Diamond | 29.29 | Normalized accuracy (%) | 99.2 |
| Open Ko-LLM Leaderboard | 42.16 | Average (%) | 84.9 |
| Open Korean LLM Leaderboard | 42.16 | Average Score (%) | 84.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 21.53 | Exact match, flexible extract (%) | 53.8 |
| Open Korean LLM - Ko-IFEval | 27.74 | Strict accuracy, prompt/instruction mean (%) | 44.7 |
| Open Korean LLM - Ko-EQ-Bench | 0 | EQ-Bench score | 24.5 |
Interactive version: theaggregate.ai/model?slug=solar-10-7b-instruct-t3q-ko-solar-dpo-ties · How It Works · Data refreshed daily, snapshot 2026-09-19.