Music Flamingo: benchmark results
Provider: Other.
Aggregate Rating 1619 ± 35, rank #846 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BASS - Artist Attribution | 56.97 | Exact-match accuracy (%) on the 64 artist attribution questi | 92.3 |
| BASS - Artist Counting | 33.54 | Exact-match accuracy (%) on the 229 artist counting question | 92.3 |
| BASS - Artist Localization | 4.69 | Exact-match accuracy (%) on the 247 artist localization ques | 84.6 |
| MusTBENCH - Local Transition Recognition | 56.3 | Local Transition Recognition (LTR, 208 questions): multiple- | 81.8 |
| HumMusQA | 58.5 | Accuracy (%) on HumMusQA's 320 expert-written four-option mu | 80 |
| MusTBENCH - Transition-Aware Description | 23.4 | Transition-Aware Description (TAD, 208 questions): the model | 72.7 |
| BASS - Artist Duration | 4.27 | Exact-match accuracy (%) on the 327 artist duration question | 69.2 |
| BASS - Single-Gene Detection | 49.25 | Exact-match accuracy (%) on the 335 single-gene detection qu | 69.2 |
| MusTBENCH - Temporal Source Grounding | 38.5 | Temporal Source Grounding (TSG, 400 questions): the model gi | 63.6 |
| BASS - Full Structural Segmentation | 13.07 | Label-matched intersection over union (%) on the 74 full str | 61.5 |
| MusTBENCH | 31.1 | Total Avg. (self-reported) | 60 |
| HalluAudio - Music | 56.3 | Average classification accuracy (%) over the nine music task | 55.6 |
Interactive version: theaggregate.ai/model?slug=music-flamingo · How It Works · Data refreshed daily, snapshot 2026-11-08.