# Full-duplex dialogue on HumDial 1.5 (dev)

Source: [Full-duplex dialogue on HumDial 1.5 (dev)](<https://www.sota2.com/research/sota/full-duplex-dialogue-on-humdial-1-5-dev>)

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

Data updated: 2026-02-26T05:45:19.526Z

Evaluation of full-duplex dialogue performance on the development split of the HumDial 1.5 dataset, focusing on response latency and interruption handling 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 | Full-duplex dialogue |
| Dataset (including any stated split) | HumDial 1.5 (dev) |
| Metric | First Response Delay |
| Unit | TIME |
| Direction | Lower 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

- [First Response Delay](<https://www.sota2.com/research/sota/full-duplex-dialogue-on-humdial-1-5-dev.md?metric=First%20Response%20Delay>): Lower is better; unit: TIME.

- [Interruption Score](<https://www.sota2.com/research/sota/full-duplex-dialogue-on-humdial-1-5-dev.md?metric=Interruption%20Score>): Higher is better; unit: PERCENT.

- [Rejection Score](<https://www.sota2.com/research/sota/full-duplex-dialogue-on-humdial-1-5-dev.md?metric=Rejection%20Score>): Higher is better; unit: PERCENT.

- [Total Delay](<https://www.sota2.com/research/sota/full-duplex-dialogue-on-humdial-1-5-dev.md?metric=Total%20Delay>): Lower is better; unit: TIME.

## Recorded results

| Method | First Response Delay | Result date | Report paper |
| --- | --- | --- | --- |
| unit-based full-duplex dialogue framework | 1.528 | Not recorded | 697ae97ea2b28e761d87cfe5 |
| baseline | 2.753 | Not recorded | 697ae97ea2b28e761d87cfe5 |

### unit-based full-duplex dialogue framework

First Response Delay: 1.528. Result ID: 699667745698665ec4ade341.

[Reporting paper](<https://www.sota2.com/research/paper/697ae97ea2b28e761d87cfe5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697ae97ea2b28e761d87cfe5.md>)

Recorded evaluation configuration: {"Decision Module":"Qwen3-Omni","ASR":"Paraformer","TTS":"IndexTTS1.5","VAD":"Silero VAD"}. This record does not assert independent reproduction.

### baseline

First Response Delay: 2.753. Result ID: 699667745698665ec4ade340.

[Reporting paper](<https://www.sota2.com/research/paper/697ae97ea2b28e761d87cfe5.md>)

Recorded evaluation configuration: {"Type":"ASR-to-LLM cascade"}. This record does not assert independent reproduction.

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