# Speech-to-Text Translation on CoVoST2 (average)

Source: [Speech-to-Text Translation on CoVoST2 (average)](<https://www.sota2.com/research/sota/speech-to-text-translation-on-covost2-average>)

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

Data updated: 2026-06-16T10:07:08.474Z

Evaluation of Speech-to-Text Translation performance on the aggregated results of the CoVoST2 dataset, measured using COMET and BLEU scores.

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 Translation |
| Dataset (including any stated split) | CoVoST2 (average) |
| Metric | Average COMET Score |
| 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

- [Average COMET Score](<https://www.sota2.com/research/sota/speech-to-text-translation-on-covost2-average.md?metric=Average%20COMET%20Score>): Higher is better; unit: SCALAR.

- [Average BLEU Score](<https://www.sota2.com/research/sota/speech-to-text-translation-on-covost2-average.md?metric=Average%20BLEU%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Average COMET Score | Result date | Report paper |
| --- | --- | --- | --- |
| SPRI-3B | 80.57 | Not recorded | 6a30d8ddd33dfb23e783cf92 |
| Qwen2.5-Omni-3B FT | 76.59 | Not recorded | 6a30d8ddd33dfb23e783cf92 |
| Qwen2.5-Omni-3B | 68.07 | Not recorded | 6a30d8ddd33dfb23e783cf92 |

### SPRI-3B

Average COMET Score: 80.57. Result ID: 6a312038fbc373592c2e0e0e.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8ddd33dfb23e783cf92.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a30d8ddd33dfb23e783cf92.md>)

Recorded evaluation configuration: {"Model Type":"MoE Upcycled"}. This record does not assert independent reproduction.

### Qwen2.5-Omni-3B FT

Average COMET Score: 76.59. Result ID: 6a312038fbc373592c2e0e0d.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8ddd33dfb23e783cf92.md>)

Recorded evaluation configuration: {"Model Type":"Dense","Training":"Fine-tuned"}. This record does not assert independent reproduction.

### Qwen2.5-Omni-3B

Average COMET Score: 68.07. Result ID: 6a312038fbc373592c2e0e0c.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8ddd33dfb23e783cf92.md>)

Recorded evaluation configuration: {"Model Type":"Dense"}. This record does not assert independent reproduction.

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