# Speech-to-text Instruction Following on AlpacaEval

Source: [Speech-to-text Instruction Following on AlpacaEval](<https://www.sota2.com/research/sota/speech-to-text-instruction-following-on-alpacaeval>)

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

Data updated: 2026-04-21T11:15:56.988Z

Evaluation of model performance on speech-to-text instruction following tasks using the AlpacaEval dataset, measured by GPTScore.

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 | Speech-to-text Instruction Following |
| Dataset (including any stated split) | AlpacaEval |
| Metric | GPTScore |
| 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

- [GPTScore](<https://www.sota2.com/research/sota/speech-to-text-instruction-following-on-alpacaeval.md?metric=GPTScore>): Higher is better; unit: SCALAR.

## Recorded results

| Method | GPTScore | Result date | Report paper |
| --- | --- | --- | --- |
| Freeze-Omni | 3.9 | Not recorded | 68fa7c551e209131cea8d9f8 |
| ELLSA | 3.09 | Not recorded | 68fa7c551e209131cea8d9f8 |
| Moshi | 1.84 | Not recorded | 68fa7c551e209131cea8d9f8 |

### Freeze-Omni

GPTScore: 3.9. Result ID: 69e75b7447f34ce924ce8ae0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c551e209131cea8d9f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa760a1e209131cea6b0d4.md>)

Recorded evaluation configuration: {"mode":"S2T"}. This record does not assert independent reproduction.

### ELLSA

GPTScore: 3.09. Result ID: 69e75b7447f34ce924ce8ae1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c551e209131cea8d9f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c551e209131cea8d9f8.md>)

Recorded evaluation configuration: {"mode":"S2T"}. This record does not assert independent reproduction.

### Moshi

GPTScore: 1.84. Result ID: 69e75b7447f34ce924ce8adf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c551e209131cea8d9f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75001e209131cea652b7.md>)

Recorded evaluation configuration: {"mode":"S2T"}. This record does not assert independent reproduction.

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