KMMAU - Word Order: leaderboard
The order in which words are spoken, a four-way choice, on native Korean speech in KMMAU (Kim et al., 2026). Accuracy in percent; higher is better.
Metric: Accuracy (%; four-choice question on the order of spoken words; native Korean audio from KSS, KMSAV and the Seoul Corpus; single run at temperature 0.7). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Raon-Speech | 98.7 | |
| 2 | Step-Audio-2-mini | 94 | |
| 3 | Qwen2.5-Omni | 80.4 | #1499 |
| 4 | Fun-Audio-Chat | 74.1 | |
| 5 | MiniCPM-o 4.5 | 54.3 | |
| 6 | Audio Flamingo 3 | 54.1 | |
| 7 | InteractiveOmni | 52.6 | |
| 8 | HyperCLOVA X 8B Omni | 15 |
Interactive version: theaggregate.ai/benchmark?slug=kmmau-word-order · How It Works · Data refreshed daily, snapshot 2026-11-12.