List your company

Speech-to-Text Translation on CoVoST2 en-zh

55.22BLEUAudio-Interaction
32.194438.172244.1550.1278Nov 14, 2023Apr 17, 2024Sep 19, 2024Feb 22, 2025Jul 27, 2025Dec 29, 2025Jun 3, 2026

Evaluation results

MethodLinks
Audio-InteractionSize=3B2026.0655.22
MiniCPM-o 4.5Size=9B2026.0449.9
Qwen3-OmniSize=30B-A3B2026.0446.6
Phi-4-multimodalSize=5.6B2026.0646.3
Qwen2-AudioSize=7B2026.0645.2
Qwen-Audiozero-shot=true2023.1141.5
Qwen2.5-OmniTrain audio duration(h)=>1000k2026.0441.4
Qwen2.5-OmniSize=7B2026.0641.4
Qwen2.5-OmniSize=3B2026.0639.5
SLAMTrain audio duration(h)=1.8k2026.0437.34
Kimi-AudioSize=9B2026.0436.6
TASUTrain audio duration(h)=02026.0433.35
TASU2Train audio duration(h)=02026.0433.08