Open Korean LLM Leaderboard Leaderboard

Open Korean LLM Leaderboard — leaderboard

Korean language LLM evaluation across Ko-GPQA, Ko-Winogrande, Ko-GSM8K, Ko-EQ Bench, Ko-IFEval, and KorNAT tasks.

Metric: Average Score (%). Source: huggingface.co. Status: saturated. 1166 models tracked.

#	Model	Score
1	Gemma-2-Ataraxy-Gemmasutra-9B-slerp	825.4
2	Gemma-2-9B-It-SPPO-Iter3	818.9
3	recoilme-gemma-2-9B-v0.4	813.12
4	Gemma-2-9B-It-SPPO-Iter1	809.48
5	Gemma-2-9B-It-SPPO-Iter2	808.28
6	Gemma 2 9B (IT)	798.03
7	Gemmasutra-9B-v1	795.55
8	Qwen2.5-14B-Gutenberg-1e-Delta	786.37
9	qwen2.5-reinstruct-alternate-lumen-14B	785.76
10	ko-gemma-2-9B-it	785.66
11	Qwen2.5-Lumen-14B	785.51
12	SuperNova-Medius	784.33
13	Qwen2.5-14B-Gutenberg-Instruct-Slerpeno	780.9
14	recoilme-gemma-2-9B-v0.2	780.58
15	gemma2-gutenberg-27B	778.27