%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Join Now
On September 23, the Qwen large model launched the Qwen-Audio-3.1 series of audio large models, forming a complete audio capability stack covering understanding, generation, interaction, and creation. The series includes speech recognition Qwen-Audio-3.1-ASR, audio understanding ASR-Next, text-to-speech TTS, audio creation TTS-Next, and real-time interaction Realtime. The API has been listed on the Qwen AI platform, and the ASR-Next API will be launched soon. At the same time, all products are being discounted, with TTS dropping by about 70%, Realtime by about 85%, and ASR by up to 95%.
In terms of capabilities, ASR enhances multilingual and dialect recognition as well as context understanding. It now supports native transcription polishing, automatically removing filler words and repeated expressions, and supports end-to-end role-based transcription, jointly outputting speaker tags, timestamps, and text; one model supports 30 languages and 16 Chinese dialects, with a first-word response time of about 160 milliseconds for streaming recognition.
ASR-Next is based on the next-generation architecture, expanding audio understanding from text in speech to emotions, ambient sounds, and mechanical sounds, supporting voice description, event localization, and audio question-answering reasoning. TTS supports natural cross-language voice color migration and controls emotion and speaking speed through instructions. TTS-Next is aimed at audio creation, unifying the generation of human voice, sound effects, and ambient sounds, supporting multi-role voice replication, fine-grained timestamps, and 48kHz output, serving scenarios such as podcasts, audiobooks, and film and game applications. Realtime realizes full-duplex interaction based on a multi-teacher distillation architecture, allowing interruptions at any time, supporting real-time language switching, emotion understanding, and tool calls during dialogue.
In testing, ASR achieved an average CER of 4.55% on open-source dialect test sets, an average CER of 10.38% on self-built Chinese dialect sets, and an average semantic sentence accuracy of 82.10% for converting dialects to Mandarin; ASR-Flash-Next and ASR-Flash ranked first in five and three out of eight metrics across four test sets, comprehensively outperforming the previous Fun-ASR cascading system.
Currently, Realtime is integrated with agents like Qoder and Qwen Office, as well as smart hardware like the Qwen AI glasses, continuing the trend of voice becoming a natural interface for human-computer interaction.
Qwen-Audio-3.1-ASR:
https://www.qianwenai.com/models/qwen-audio-3.1-asr-flash
Qwen-Audio-3.1-TTS:
https://www.qianwenai.com/models/qwen-audio-3.1-tts-flash
Qwen-Audio-3.1-TTS-Next:
https://www.qianwenai.com/models/qwen-audio-3.1-tts-next
Qwen-Audio-3.1-Realtime:
https://www.qianwenai.com/models/qwen-audio-3.1-realtime-plus
(The Qwen-Audio-3.1-ASR-Next API will be launched soon)