BASS - Section Structural Lyric Transcription: leaderboard
Metric: Word error rate (a ratio, not a percentage; 0 is perfect and insertions leave it unbounded above, averaged over matched sections) for transcribing the lyrics of one section type or one named section instance against Genius reference lyrics; audio language models get a text prompt (one of ten paraphrases) and a full commercial music recording from BASS's songs; free-form answers; lower is better. Source: arxiv.org. Saturation forecast: Around October 2026. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 2.4 | #145 |
| 2 | Voxtral-Small-24B-2507 | 15.46 | #565 |
| 3 | SALMONN-7B | 20.59 | #1274 |
| 4 | Phi-4 Multimodal Instruct | 364.54 | #896 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=bass-section-structural-lyric-transcription · How It Works · Data refreshed daily, snapshot 2026-10-11.