KoLlama-3-8B-Instruct: benchmark results

Provider: Other. Access: Open.

Aggregate Rating 1459 ± 43, rank #3451 of 10552 rated models, from 9 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open Korean LLM - Ko-IFEval45.77Strict accuracy, prompt/instruction mean (%)90.3
Open Korean LLM - Ko-GSM8K (flexible extract)55.57Exact match, flexible extract (%)89.5
Open Korean LLM - Ko-GPQA-Diamond26.26Normalized accuracy (%)85.5
Open Korean LLM - Ko-EQ-Bench33.44EQ-Bench score83.2
Open Korean LLM Leaderboard40.05Average Score (%)75.2
Open Korean LLM - KorNAT-CKA25.25Normalized accuracy (%)56.3
Open Korean LLM - KorNAT-Helpful46.85Normalized accuracy (%)50
Open Korean LLM - Ko-Winogrande58.33Accuracy (%)37.8
Open Korean LLM - KorNAT-Harmless62.07Normalized accuracy (%)37.5

Interactive version: theaggregate.ai/model?slug=kollama-3-8b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-16.