# English-to-Chinese Speech-to-Speech Conversation on InstructS2S Eval

Source: [English-to-Chinese Speech-to-Speech Conversation on InstructS2S Eval](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval>)

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

Data updated: 2026-04-14T11:19:33.462Z

Evaluation of models on the InstructS2S Eval dataset for English-to-Chinese Speech-to-Speech Conversation, utilizing GPT scores, ASR Error Rate (ASR-ER), and Off-Target metrics.

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 | English-to-Chinese Speech-to-Speech Conversation |
| Dataset (including any stated split) | InstructS2S Eval |
| Metric | GPT Score (T) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [GPT Score (T)](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval.md?metric=GPT%20Score%20(T)>): Higher is better; unit: SCALAR.

- [GPT Score (S)](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval.md?metric=GPT%20Score%20(S)>): Higher is better; unit: SCALAR.

- [ASR Error Rate](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval.md?metric=ASR%20Error%20Rate>): Lower is better; unit: PERCENT.

- [Off-Target Rate (T)](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval.md?metric=Off-Target%20Rate%20(T)>): Lower is better; unit: PERCENT.

- [Off-Target Rate (S)](<https://www.sota2.com/research/sota/english-to-chinese-speech-to-speech-conversation-on-instructs2s-eval.md?metric=Off-Target%20Rate%20(S)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | GPT Score (T) | Result date | Report paper |
| --- | --- | --- | --- |
| GLM-4-Voice | 4.29 | Not recorded | 69ddca6a7d0172140ec79eb2 |
| CSLM | 3.31 | Not recorded | 69ddca6a7d0172140ec79eb2 |

### GLM-4-Voice

GPT Score (T): 4.29. Result ID: 69de22ba47f34ce924c68e78.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca6a7d0172140ec79eb2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75f91e209131cea6abba.md>)

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

### CSLM

GPT Score (T): 3.31. Result ID: 69de22ba47f34ce924c68e79.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca6a7d0172140ec79eb2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ddca6a7d0172140ec79eb2.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.