qwen-audio-3.0-realtime-plus

Description

Alibaba Qwen-Audio-3.0-Realtime-Plus — full-duplex speech-to-speech conversation over a WebSocket session, which the vendor reports as first overall in the Artificial Analysis Speech-to-Speech ranking. The standard tier favours answer quality, keeping speech reasoning intact while parallel inference and end-to-end streaming hold response latency low. Reached over GET /v1/realtime. Audio input and audio output are billed on their own faces, well above the text rate.