# Music Reasoning on Human Evaluation Set 1.0 (test)

Source: [Music Reasoning on Human Evaluation Set 1.0 (test)](<https://www.sota2.com/research/sota/music-reasoning-on-human-evaluation-set-1-0-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:39:53.652Z

Human evaluation of model performance on music reasoning tasks, assessed using subjective metrics such as Correctness, Musicality, and Clarity.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Music Reasoning |
| Dataset (including any stated split) | Human Evaluation set for Music Reasoning 1.0 (test) |
| Metric | Correctness |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Correctness](<https://www.sota2.com/research/sota/music-reasoning-on-human-evaluation-set-1-0-test.md?metric=Correctness>): Higher is better; unit: SCALAR.

- [Musicality](<https://www.sota2.com/research/sota/music-reasoning-on-human-evaluation-set-1-0-test.md?metric=Musicality>): Higher is better; unit: SCALAR.

- [Clarity](<https://www.sota2.com/research/sota/music-reasoning-on-human-evaluation-set-1-0-test.md?metric=Clarity>): Higher is better; unit: SCALAR.

- [Overall Score](<https://www.sota2.com/research/sota/music-reasoning-on-human-evaluation-set-1-0-test.md?metric=Overall%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Correctness | Result date | Report paper |
| --- | --- | --- | --- |
| MuseChat | 4.63 | Not recorded | 68fa72711e209131cea5659e |
| Vicuna-7B | 3.07 | Not recorded | 68fa72711e209131cea5659e |
| Vicuna w/ Music | 1.24 | Not recorded | 68fa72711e209131cea5659e |

### MuseChat

Correctness: 4.63. Result ID: 699238fe3a6c3148490d6c25.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72711e209131cea5659e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea5659e.md>)

Recorded evaluation configuration: {"Input Modality":"Music Title + Embeddings"}. This record does not assert independent reproduction.

### Vicuna-7B

Correctness: 3.07. Result ID: 699238fe3a6c3148490d6c23.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72711e209131cea5659e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71aa1e209131cea526df.md>)

[Implementation](<https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge>)

Recorded evaluation configuration: {"Input Modality":"Music Title"}. This record does not assert independent reproduction.

### Vicuna w/ Music

Correctness: 1.24. Result ID: 699238fe3a6c3148490d6c24.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72711e209131cea5659e.md>)

Recorded evaluation configuration: {"Input Modality":"Music Embeddings"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
