MusicBench-Temporal: leaderboard

Temporal music understanding from audio (MusicBench, You et al., 2026, introduced with the GaMMA model): 998 multiple-choice questions written by music experts that require locating a stated time span of a song and identifying its chords, structural section, lyrics, instruments or vocals. Accuracy in percent, higher is better.

Metric: Accuracy (%; 998 expert-written multiple-choice questions on chords, structure, lyrics, instruments and vocals within a stated time span of a song, answered from the audio). Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked.

Top models

#ModelScoreOverall rank
1Gemini 2.5 Pro74.6#262

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=musicbench-temporal · How It Works · Data refreshed daily, snapshot 2026-11-12.