Speech-to-Text Translation on FLEURS 9×27 directions (test)
78.7araMCAT-Large-27B
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| MCAT-Large-27BModel Category=End-to-end S2TT Models2025.12 | 78.7 | 80.3 | 86.3 | 83.2 | 79.8 | 82.4 | 78 | 83.2 | 79.5 | 81.3 | |
| Whisper-Large-V3 + NLLB-200-3.3BModel Category=Cascaded ASR+MT Models2025.12 | 78.1 | 79.8 | 83.4 | 81.6 | 79.9 | 81.2 | 78.2 | 82.3 | 78.3 | 80.3 | |
| MCAT-Small-9BModel Category=End-to-end S2TT Models2025.12 | 77.8 | 79.8 | 85.9 | 82.3 | 79.7 | 81.9 | 78.6 | 82.1 | 79.3 | 80.8 | |
| Whisper-Large-V3 + LLaMAX3-8B-AlpacaModel Category=Cascaded ASR+MT Models2025.12 | 76 | 78.6 | 81.3 | 79.4 | 78.4 | 79.3 | 77 | 79.2 | 77.1 | 78.5 | |
| SeamlessM4T-V2-LargeModel Category=End-to-end S2TT Models2025.12 | 69.4 | 72.6 | 84.3 | 71.2 | 69.1 | 73.5 | 68.6 | 71.2 | 71.9 | 72.4 | |
| Whisper-Large-V3 + NLLB-200-3.3BArchitecture=Cascaded ASR+MT2025.12 | 21.5 | 18.9 | 30.2 | 24.2 | 19.4 | 19.9 | 17.5 | 23.7 | 18.9 | 21.6 | |
| MCAT-Small-9BArchitecture=End-to-end S2TT2025.12 | 20.2 | 18.4 | 32.7 | 24.5 | 18.9 | 21.4 | 16.8 | 23.5 | 18.8 | 21.7 | |
| MCAT-Large-27BArchitecture=End-to-end S2TT2025.12 | 20.2 | 17.9 | 31.7 | 24.4 | 18.3 | 21.1 | 15.7 | 24.2 | 18.2 | 21.3 | |
| Whisper-Large-V3 + LLaMAX3-8B-AlpacaArchitecture=Cascaded ASR+MT2025.12 | 17.5 | 16.1 | 25.2 | 20.6 | 16.1 | 16.9 | 14.5 | 19 | 16.2 | 18 | |
| SeamlessM4T-V2-LargeArchitecture=End-to-end S2TT2025.12 | 15.5 | 13.2 | 30.9 | 15.3 | 12.2 | 14.4 | 11.5 | 15.7 | 13.9 | 15.8 |