# Co-speech Gesture Generation on BEAT 2

Source: [Co-speech Gesture Generation on BEAT 2](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat-2>)

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

Data updated: 2026-02-26T05:53:34.212Z

Evaluation of Co-speech Gesture Generation performance on the BEAT 2 dataset, assessed using human evaluation metrics for naturalness, synchrony, and emotion.

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 | Co-speech Gesture Generation |
| Dataset (including any stated split) | BEAT 2 |
| Metric | Naturalness |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Naturalness](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat-2.md?metric=Naturalness>): Higher is better; unit: PERCENT.

- [Synchrony](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat-2.md?metric=Synchrony>): Higher is better; unit: PERCENT.

- [Emotion](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat-2.md?metric=Emotion>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Naturalness | Result date | Report paper |
| --- | --- | --- | --- |
| 3DGesPolicy | 91.67 | Not recorded | 697846daa2b28e761d87c7e3 |
| 3DGesPolicy | 91.67 | Not recorded | 697846daa2b28e761d87c7e3 |
| 3DGesPolicy | 83.33 | Not recorded | 697846daa2b28e761d87c7e3 |
| GT | 75 | Not recorded | 697846daa2b28e761d87c7e3 |
| 3DGesPolicy | 25 | Not recorded | 697846daa2b28e761d87c7e3 |
| EMAGE | 16.67 | Not recorded | 697846daa2b28e761d87c7e3 |
| DiffSHEG | 8.33 | Not recorded | 697846daa2b28e761d87c7e3 |
| MambaTalk | 8.33 | Not recorded | 697846daa2b28e761d87c7e3 |

### 3DGesPolicy

Naturalness: 91.67. Result ID: 69978c945698665ec4e1ae8c.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. DiffSHEG \[8\]"}. This record does not assert independent reproduction.

### 3DGesPolicy

Naturalness: 91.67. Result ID: 69978c945698665ec4e1ae90.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. MambaTalk \[9\]"}. This record does not assert independent reproduction.

### 3DGesPolicy

Naturalness: 83.33. Result ID: 69978c945698665ec4e1ae8e.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. EMAGE \[4\]"}. This record does not assert independent reproduction.

### GT

Naturalness: 75. Result ID: 69978c945698665ec4e1ae93.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. GT"}. This record does not assert independent reproduction.

### 3DGesPolicy

Naturalness: 25. Result ID: 69978c945698665ec4e1ae92.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. GT"}. This record does not assert independent reproduction.

### EMAGE

Naturalness: 16.67. Result ID: 69978c945698665ec4e1ae8f.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c11e209131cea5867d.md>)

[Implementation](<https://pantomatrix.github.io/EMAGE/>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. EMAGE \[4\]"}. This record does not assert independent reproduction.

### DiffSHEG

Naturalness: 8.33. Result ID: 69978c945698665ec4e1ae8d.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72cb1e209131cea5876b.md>)

[Implementation](<https://jeremycjm.github.io/proj/DiffSHEG>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. DiffSHEG \[8\]"}. This record does not assert independent reproduction.

### MambaTalk

Naturalness: 8.33. Result ID: 69978c945698665ec4e1ae91.

[Reporting paper](<https://www.sota2.com/research/paper/697846daa2b28e761d87c7e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79811e209131cea7dddf.md>)

Recorded evaluation configuration: {"Comparison Group":"Ours vs. MambaTalk \[9\]"}. This record does not assert independent reproduction.

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