AudioBench - VoxCeleb Accent: leaderboard
AudioBench VoxCeleb Accent slice measuring speaker-accent classification from audio inputs.
Metric: Mean Score (%). Source: github.com. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen-Audio-Chat | 48.05 | #1244 |
| 2 | SALMONN-7B | 34.22 | #1293 |
| 3 | Phi-4 Multimodal Instruct | 26.82 | #906 |
| 4 | Qwen2-Audio-7B-Instruct | 25.93 | #880 |
| 5 | SeaLLMs-Audio-7B | 8.66 | #1018 |
Interactive version: theaggregate.ai/benchmark?slug=audiobench-voxceleb-accent · How It Works · Data refreshed daily, snapshot 2026-10-20.