# Speech Generation on Long-Audio benchmark English

Source: [Speech Generation on Long-Audio benchmark English](<https://www.sota2.com/research/sota/speech-generation-on-long-audio-benchmark-english>)

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

Data updated: 2026-03-11T06:52:29.078Z

Evaluation of Speech Generation performance on the 'Long-Audio benchmark English' dataset, assessed using metrics including WER, SIM-Mean, and SIM-Std.

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 Generation |
| Dataset (including any stated split) | Long-Audio benchmark English |
| Metric | WER |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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

- [WER](<https://www.sota2.com/research/sota/speech-generation-on-long-audio-benchmark-english.md?metric=WER>): Lower is better; unit: PERCENT.

- [SIM-Mean](<https://www.sota2.com/research/sota/speech-generation-on-long-audio-benchmark-english.md?metric=SIM-Mean>): Higher is better; unit: SCALAR.

- [SIM-Std](<https://www.sota2.com/research/sota/speech-generation-on-long-audio-benchmark-english.md?metric=SIM-Std>): Lower is better; unit: SCALAR.

## Recorded results

| Method | WER | Result date | Report paper |
| --- | --- | --- | --- |
| Fish Audio S2 | 4.38 | Not recorded | 69b0f6e92d603e6de3e45b93 |
| Fish Audio S1 | 6.26 | Not recorded | 69b0f6e92d603e6de3e45b93 |
| Qwen3-TTS | 7.69 | Not recorded | 69b0f6e92d603e6de3e45b93 |
| VibeVoice | 28 | Not recorded | 69b0f6e92d603e6de3e45b93 |

### Fish Audio S2

WER: 4.38. Result ID: 69b1108c6dbb7075b95ad785.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f6e92d603e6de3e45b93.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f6e92d603e6de3e45b93.md>)

[Implementation](<https://github.com/fishaudio/fish-speech>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Fish Audio S1

WER: 6.26. Result ID: 69b1108c6dbb7075b95ad786.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f6e92d603e6de3e45b93.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75991e209131cea68dc0.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Qwen3-TTS

WER: 7.69. Result ID: 69b1108c6dbb7075b95ad787.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f6e92d603e6de3e45b93.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6973006ab2be0ed85a78c8c5.md>)

[Implementation](<https://github.com/QwenLM/Qwen3-TTS>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### VibeVoice

WER: 28. Result ID: 69b1108c6dbb7075b95ad788.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f6e92d603e6de3e45b93.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7adc1e209131cea858bb.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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