UniVerseBench - Korean: leaderboard
Metric: Accuracy (%; questions on Korean music of UniVerseBench, multiple-choice questions on folk and traditional music recordings from 372 recordings in over 38 languages, 5,042 questions in total, verified by music experts; direct option generation). Source: arxiv.org. Saturation forecast: Around December 2026. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash | 77.6 |
| 2 | Qwen 3.5 Omni Plus | 75.5 |
Interactive version: theaggregate.ai/benchmark?slug=universebench-korean · How It Works · Data refreshed daily, snapshot 2026-09-26.