EEVE-Korean-10.8B-RAFT: benchmark results
Provider: Yanolja. Access: Open.
Unified ELO 1522 ± 17, rank #976 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-CKA | 33.29 | Normalized accuracy (%) | 91.5 |
| Open Korean LLM Leaderboard | 43.86 | Average Score (%) | 89.3 |
| Open Ko-LLM Leaderboard | 43.86 | Average (%) | 89.2 |
| Open Korean LLM - Ko-EQ-Bench | 35.71 | EQ-Bench score | 85.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 49.81 | Exact match, flexible extract (%) | 83.8 |
| Open Korean LLM - Ko-Winogrande | 64.72 | Accuracy (%) | 80.3 |
| Open Korean LLM - Ko-IFEval | 33.04 | Strict accuracy, prompt/instruction mean (%) | 64.8 |
| Open Korean LLM - Ko-GPQA-Diamond | 22.73 | Normalized accuracy (%) | 40.9 |
| Open Korean LLM - KorNAT-Helpful | 44 | Normalized accuracy (%) | 27 |
| Open Korean LLM - KorNAT-Harmless | 61.24 | Normalized accuracy (%) | 25.8 |
Interactive version: theaggregate.ai/model?slug=eeve-korean-10-8b-raft · How It Works · Data refreshed daily, snapshot 2026-09-19.