KVoiceBench: leaderboard

Metric: Mean of the nine sub-benchmark scores (each on 0-100 under its source metric: accuracy, GPT-5.4 judge score, instruction-following accuracy or refusal rate; 7,306 Korean spoken samples transferred from VoiceBench; single run at temperature 0.7). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 8 models tracked.

Top models

#ModelScore
1MiniCPM-o-4.539.5

Interactive version: theaggregate.ai/benchmark?slug=kvoicebench · How It Works · Data refreshed daily, snapshot 2026-09-26.