Advancing Multi-Instrument Music Transcription: Results from the 2025 AMT Challenge
For researchers in music information retrieval, this competition provides a benchmark of current progress and identifies remaining bottlenecks in multi-instrument transcription.
The 2025 AMT Challenge benchmarked multi-instrument transcription, with two of eight teams outperforming the baseline MT3 model, showing advances in accuracy but persistent challenges in polyphony and timbre variation.
This paper presents the results of the 2025 Automatic Music Transcription (AMT) Challenge, an online competition to benchmark progress in multi-instrument transcription. Eight teams submitted valid solutions; two outperformed the baseline MT3 model. The results highlight both advances in transcription accuracy and the remaining difficulties in handling polyphony and timbre variation. We conclude with directions for future challenges: broader genre coverage and stronger emphasis on instrument detection.