Metraipolis › archive › dc_654b620c2191

PianoBench: Evaluating Musical Perception and Reasoning in Language Models — arXiv:2611.03839

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.03839
License (as archived)
CC0-1.0
Archived
2026-11-04T21:39:40.023Z
SHA-256 of the archived text
aa357d454942561d6d0cc00104f8358ea26b638e027e0c77738bc610e82dbaad
Stable URL of this page
https://metraipolis.com/documento/dc_654b620c2191

How to cite

PianoBench: Evaluating Musical Perception and Reasoning in Language Models — arXiv:2611.03839 — arXiv (en), https://arxiv.org/abs/2611.03839 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_654b620c2191 (Metraipolis archive dc_654b620c2191, sha256 aa357d454942561d6d0cc00104f8358ea26b638e027e0c77738bc610e82dbaad)
BibTeX
@misc{dc654b620c2191,
  title = {PianoBench: Evaluating Musical Perception and Reasoning in Language Models — arXiv:2611.03839},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03839},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_654b620c2191; sha256 aa357d454942561d6d0cc00104f8358ea26b638e027e0c77738bc610e82dbaad},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03839. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multimodal language models continue to improve on broad music understanding benchmarks, but which musical features can they perceive and reason about? We investigate this question by introducing PianoBench, a benchmark that tasks language models with reasoning about musical performances across congruent acoustic and symbolic input representations. PianoBench is derived from our new multimodal, training-scale dataset comprising more than 20,000 hours of interleaved transcripts of music and spoken musical commentary from instructional piano videos. Open-weight audio-language models achieve at most 52.8% accuracy on PianoBench, and at best only modestly outperform their accuracy on a no-audio control. In contrast, frontier models using chain of thought achieve up to 93.2% accuracy when provided transcripts of music as text, exposing a gap between perceiving musical features from audio and reasoning about musical events when they are given explicitly. We then test whether mid-training on interleaved musical instruction can improve question answering without chain of thought, raising Qwen 3.5 9B Base's zero-shot accuracy on PianoBench with music transcripts from 42.8% to 56.0% after instruction tuning. These experiments suggest that instructional videos may represent an overlooked source of multimodal supervision for music. Available: this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.03839 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T21:35:49.623Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).