Model Studio speech recognition transcribes live audio and recordings, with multilingual and dialect support, hotwords, context, text refinement, and speaker diarization.
Features
Speech recognition converts speech in audio or video into text for voice input, live captions, meeting records, and call analysis. Model Studio supports both continuously streamed audio and complete recordings. Language, dialect, and enhancement capabilities vary by model.
Choose a model
Choose a model based on the audio input, use case, and required result delivery. The following recommendations use the Qwen-Audio 3.1 series.
| Type | Use case | Recommended model | Key difference |
|---|---|---|---|
Real-time | Continuous transcription, such as live captions or meetings | Continuously sends audio while receiving intermediate and final results. | |
Short spoken input, such as dictation or voice messages | Returns final sentence results by default; intermediate results and text refinement can be enabled separately. | ||
Non-real-time | Short recordings of up to 5 minutes | Returns recognition results in the current request, with optional SSE output. | |
Long recordings and background audio/video transcription | Submits an asynchronous task; download transcription files when it completes. |
Recognition methods
| Method | Audio input | Result delivery |
|---|---|---|
Continuously sends audio chunks to display transcription while speech is being captured. | Receives results while sending audio. Delivery of intermediate and final results varies by model. | |
Submits a complete audio or video file to process an existing recording. | Synchronous calls return results in the current request, optionally as SSE events where supported. Asynchronous calls return a task ID first, then provide result files after completion. |
Improve recognition and use results
After selecting a model, use its supported features to improve specific-word recognition, text readability, and structured results. For configuration details, see the real-time speech recognition guide or non-real-time speech recognition guide.
-
Hotwords and context: Hotwords provide the intended spelling of names, brands, and other terms. Context adds earlier dialogue and business background. For configuration, model coverage, and tuning, see Improve recognition accuracy.
-
Text refinement: Organizes fillers, repetitions, and self-corrections in speech. Native text refinement is available in qwen-audio-3.1-asr-flash-message and qwen-audio-3.1-asr-flash.
-
Preserve dialect expressions: qwen-audio-3.1-asr-flash-streaming, qwen-audio-3.1-asr-flash-message, qwen-audio-3.1-asr-flash, qwen-audio-3.1-asr-flash-filetrans can preserve dialect-specific expressions. Recognizing a dialect and preserving its written expressions are separate capabilities.
-
Speaker diarization: Labels speakers in transcription results to distinguish participants in meetings or interviews. qwen-audio-3.1-asr-flash and qwen-audio-3.1-asr-flash-filetrans support this feature.
-
Timestamps: Associates sentences or words with positions in audio for captions, keyword highlighting, and seeking.
-
Sensitive-word filtering: Replaces specified words with asterisks or removes them from transcription text.
For audio preparation, connection and concurrency management, asynchronous jobs, and error handling, see Speech recognition best practices.
Models and languages
The following lists show the languages and dialects supported by each model family and version.
Qwen-Audio-ASR-Streaming
- qwen-audio-3.1-asr-flash-streaming, qwen-audio-3.0-asr-flash-streaming: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
Qwen-Audio-ASR-Message
- qwen-audio-3.1-asr-flash-message: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
Qwen-Audio-ASR
- qwen-audio-3.1-asr-flash: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak, Turkish, Persian, Marathi, Malayalam, Nepali, Kyrgyz, Azerbaijani, Georgian, Armenian, Slovenian, Albanian, Estonian, Latvian, Lithuanian, Catalan, Basque, Galician, Odia, Divehi, Assamese, Kurdish, Tigre, Western Frisian, Bashkir, Belarusian, Tatar, Interlingua, Eastern Mari, Toki Pona
- qwen-audio-3.0-asr-flash: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
Qwen-Audio-ASR-Filetrans
- qwen-audio-3.1-asr-flash-filetrans, qwen-audio-3.0-asr-flash-filetrans: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
Fun-ASR
- fun-asr-realtime: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
- fun-asr-realtime-2026-02-28: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese
- fun-asr-flash-8k-realtime, fun-asr-flash-8k-realtime-2026-01-28: Chinese
- fun-asr: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
- fun-asr-flash-2026-06-15: Chinese (Mandarin, Cantonese, Wu, Hokkien, Hakka, Gan, Xiang, Jin; plus regional accents including Central Plains, Southwestern, Ji-Lu, Jianghuai, Lan-Yin, Jiao-Liao, Northeastern, Beijing, and Hong Kong/Taiwan—covering Henan, Shaanxi, Hubei, Sichuan, Chongqing, Yunnan, Guizhou, Guangdong, Guangxi, Hebei, Tianjin, Shandong, Anhui, Nanjing, Jiangsu, Hangzhou, Gansu, and Ningxia accents), English, Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Filipino, Hindi, Arabic, French, German, Spanish, Portuguese, Russian, Italian, Dutch, Swedish, Danish, Finnish, Norwegian, Greek, Polish, Czech, Hungarian, Romanian, Bulgarian, Croatian, Slovak
Qwen3-ASR
- qwen3-asr-flash-realtime, qwen3-asr-flash-filetrans, qwen3-asr-flash: Chinese (Mandarin, Sichuanese, Hokkien, Wu, Cantonese), English, Japanese, German, Korean, Russian, French, Portuguese, Arabic, Italian, Spanish, Hindi, Indonesian, Thai, Turkish, Ukrainian, Vietnamese, Czech, Danish, Filipino, Finnish, Icelandic, Malay, Norwegian, Polish, Swedish.
Paraformer
- paraformer-realtime-v2, paraformer-v2: Chinese (Mandarin, Cantonese, Wu, Hokkien, Northeastern, Gansu, Guizhou, Henan, Hubei, Hunan, Ningxia, Shanxi, Shaanxi, Shandong, Sichuan, Tianjin, Jiangxi, Yunnan, Shanghai), English, Japanese, Korean, German, French, Russian
- paraformer-realtime-8k-v2, paraformer-8k-v2: Chinese (Mandarin)
Audio input requirements
For audio formats, sample rates, channels, file size, and duration, see the limits in the real-time speech recognition guide or non-real-time speech recognition guide.