AIQwen-Audio-3.1 launches with five models and voice AI price cuts
A user sharing a Yicai report says prices fell by about 70% for text-to-speech, about 85% for Realtime and up to about 95% for speech recognition.
TLDR
A user sharing a Yicai report says Qwen-Audio-3.1 launched on September 23, 2026, with five models: ASR, ASR-Next, TTS, TTS-Next and Realtime. They describe a lineup covering speech recognition and understanding, speech synthesis, audio creation and full-duplex—simultaneous two-way—interaction. The September 25 post says APIs were available except for ASR-Next, which was described as coming soon.
Qwen-Audio-3.1 launches with five models and voice AI price cuts
A user sharing a Yicai report says prices fell by about 70% for text-to-speech, about 85% for Realtime and up to about 95% for speech recognition.
TLDR
A user sharing a Yicai report says Qwen-Audio-3.1 launched on September 23, 2026, with five models: ASR, ASR-Next, TTS, TTS-Next and Realtime. They describe a lineup covering speech recognition and understanding, speech synthesis, audio creation and full-duplex—simultaneous two-way—interaction. The September 25 post says APIs were available except for ASR-Next, which was described as coming soon.