Speaker-Attributed Automatic Speech Recognition on Podcast (test)
4.46CERMOSS Transcribe Diarize
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MOSS Transcribe Diarize2026.01 | 4.46 | 6.97 | 2.5 | |
| Gemini 2.5 Pro2026.01 | 7.38 | 10.23 | 2.85 | |
| DoubaoModel=Speech Recognition Model2026.01 | 7.93 | 10.54 | 2.61 | |
| ElevenLabsModel=Scribe v12026.01 | 8.5 | 11.34 | 2.85 |