Speech-to-text overview

Choose the right model for real-time or file-based speech recognition.

Migrate from closed-source models

Map your current Whisper, Deepgram, or Google ASR service to a Bailian equivalent.

Use case	Closed-source examples	Bailian recommendation
Real-time recognition	Deepgram Nova-3, Google Chirp 3	Input methods and voice messages：qwen-audio-3.1-asr-flash-message Continuous transcription：qwen-audio-3.1-asr-flash-streaming
Offline / file transcription	OpenAI gpt-4o-transcribe, Whisper	qwen-audio-3.1-asr-flash-filetrans, qwen-audio-3.1-asr-flash

Use the four "Selection criteria"—real-time vs. offline, domain terminology, speaker diarization, and emotion recognition—to identify the right model. Then check "Recommended models" for top picks per scenario, "All models" for the full version list by model family, and "Audio specifications" for input constraints. Supported languages (including dialects) are listed under each model family in "All models".

Selection criteria

Evaluate four dimensions to find the right model.

Real-time or offline?

Real-time recognition returns results as the user speaks. Offline recognition transcribes a pre-recorded file after recording ends.

• Real-time recognition streams audio over WebSocket. Choose by use case:

  • Short voice input, such as input methods and voice messages: use qwen-audio-3.1-asr-flash-message for low first-text and overall latency, optional text polishing, and a revised final result when a speech segment ends.
  • Continuous transcription, such as meetings and classroom notes: use qwen-audio-3.1-asr-flash-streaming when users read results as they arrive and intermediate-result accuracy is a priority.

• Offline (file transcription): Uses an HTTP API to submit audio files and retrieve transcription results. Suited for call center recordings, podcasts, and interviews. Recommended model: qwen-audio-3.1-asr-flash-filetrans (hot words, prompt context, speaker diarization).

For client integrations that require stable latency, resilience on weak networks, and full-duplex noise suppression and echo cancellation, see the AOQ protocol comparison and supported models in Realtime API overview.

For HTTP, WebSocket, and SDK integration options, see Speech recognition API reference. Refer to the corresponding SDK documentation for supported models and parameters.

For real-time integration, see Real-time speech recognition. For offline integration, see Non-real-time speech recognition.

Domain terminology

Two approaches, ranked by flexibility:

1. Prompt context injection: Describe your domain in the system prompt. No setup required—the model adapts per request. Recommended models: qwen-audio-3.1-asr-flash-streaming, qwen-audio-3.1-asr-flash-filetrans, and qwen-audio-3.1-asr-flash.
2. Hot words: Supply a weighted vocabulary list. Recommended models: qwen-audio-3.1-asr-flash-streaming, qwen-audio-3.1-asr-flash-filetrans, and qwen-audio-3.1-asr-flash.

qwen-audio-3.1-asr-flash-message supports instant hot words, precompiled hot words, and conversation context. Set vocabulary or vocabulary_id for hot words, and pass previous user transcriptions and assistant replies through input.context. For parameter definitions, see Client events.

Speaker diarization

qwen-audio-3.1-asr-flash, qwen-audio-3.1-asr-flash-filetrans, qwen-audio-3.0-asr-flash-filetrans, and Fun-ASR offline models (fun-asr) support speaker diarization. To identify who said what, use qwen-audio-3.1-asr-flash-filetrans.

Emotion recognition

Qwen-ASR models detect emotions alongside transcription. Recommended models: qwen3-asr-flash-realtime (real-time) or qwen3-asr-flash-filetrans (offline).

Recommended models

Top picks for common scenarios.

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen-audio-3.1-asr-flash-message	Real-time	WebSocket	Hot words, Context	Unsupported	Unsupported	Multilingual with dialects	—
qwen-audio-3.1-asr-flash-streaming	Real-time	WebSocket	Hot words, prompt context	Unsupported	Unsupported	Multilingual with dialects	Unlimited
qwen-audio-3.1-asr-flash-filetrans	Offline	HTTP	Hot words, prompt context	Unsupported	Supported	Multilingual with dialects	12 hours / 2 GB

All models

Qwen-Audio-ASR-Message

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen-audio-3.1-asr-flash-message	Real-time	WebSocket	Hot words, Context	Unsupported	Unsupported	Multilingual with dialects	—

Supported languages:

• 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-Streaming

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen-audio-3.1-asr-flash-streaming	Real-time	WebSocket	Hot words, prompt context	Unsupported	Unsupported	Multilingual with dialects	Unlimited
qwen-audio-3.0-asr-flash-streaming	Real-time	WebSocket	Hot words, prompt context	Unsupported	Unsupported	Multilingual with dialects	Unlimited

Supported languages (by version):

• 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-Filetrans

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen-audio-3.1-asr-flash-filetrans	Offline	HTTP	Hot words, prompt context	Unsupported	Supported	Multilingual with dialects	12 hours / 2 GB
qwen-audio-3.0-asr-flash-filetrans	Offline	HTTP	Hot words, prompt context	Unsupported	Supported	Multilingual with dialects	12 hours / 2 GB

Supported languages (by version):

• 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

Qwen-Audio-ASR

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen-audio-3.1-asr-flash	Offline	HTTP	Hot words, prompt context	Unsupported	Supported	Multilingual with dialects	5 minutes / 2 GB
qwen-audio-3.0-asr-flash	Offline	HTTP	Hot words, prompt context	Unsupported	Unsupported	Multilingual with dialects	5 minutes / 2 GB

Supported languages (by version):

• qwen-audio-3.1-asr-flash, 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

Fun-ASR

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
fun-asr-realtime	Real-time	WebSocket	Hot words	Unsupported	Unsupported	Multilingual with dialects	Unlimited
fun-asr-realtime-2026-02-28	Real-time	WebSocket	Hot words	Unsupported	Unsupported	Chinese, English, Japanese, and dialects	Unlimited
fun-asr-flash-8k-realtime	Real-time	WebSocket	Hot words	Unsupported	Unsupported	Chinese	Unlimited
fun-asr-flash-8k-realtime-2026-01-28	Real-time	WebSocket	Hot words	Unsupported	Unsupported	Chinese	Unlimited
fun-asr	Offline	HTTP	Hot words	Unsupported	Supported	Multilingual with dialects	12 hours / 2 GB
fun-asr-flash-2026-06-15	Offline	HTTP	Prompt context	Unsupported	Unsupported	Multilingual with dialects	5 minutes / 2 GB

Supported languages (by version):

• Fun-ASR-Realtime main version (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-Realtime(8K) (fun-asr-flash-8k-realtime, fun-asr-flash-8k-realtime-2026-01-28): Chinese
• Fun-ASR main versions (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 (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

Qwen-ASR

Model ID	Mode	API	Accuracy boost	Emotion recognition	Speaker diarization	Languages	Max audio duration/size
qwen3-asr-flash-realtime	Real-time	WebSocket	Unsupported	Supported	Unsupported	Multilingual with dialects	Unlimited
qwen3-asr-flash-filetrans	Offline	HTTP	Unsupported	Supported	Unsupported	Multilingual with dialects	12 hours / 2 GB
qwen3-asr-flash	Offline	HTTP (OpenAI-compatible)	Unsupported	Supported	Unsupported	Multilingual with dialects	5 minutes / 10 MB

Supported languages: All Qwen-ASR models (qwen3-asr-flash-realtime, qwen3-asr-flash-filetrans, qwen3-asr-flash) share the same language list: 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 is an older-generation ASR model family. Migrate to Fun-ASR or Qwen-ASR when possible.

Model ID	API	Description
paraformer-realtime-v2	WebSocket	Real-time recognition; Chinese, English, Japanese, Korean, German, French, Russian
paraformer-realtime-8k-v2	WebSocket	Real-time recognition for 8 kHz telephony; Chinese
paraformer-v2	HTTP	File transcription with speaker diarization; Chinese, English, Japanese, Korean, German, French, Russian
paraformer-8k-v2	HTTP	File transcription for 8 kHz telephony; Chinese

Supported languages (by version):

• 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 specifications

Audio specifications for real-time and offline modes. For supported languages, see the model family subsection under "All models".

Real-time

Model ID	Input method	Audio format	Sample rate	Size/duration
qwen-audio-3.1-asr-flash-message	Binary stream	pcm, wav, mp3, opus, speex, aac, amr	16 kHz	—
qwen-audio-3.1-asr-flash-streaming, qwen-audio-3.0-asr-flash-streaming	Binary stream	pcm, wav, mp3, opus, speex, aac, amr	Any	Unlimited
Fun-ASR-Realtime series (fun-asr-realtime, etc.)	Binary stream	pcm, wav, mp3, opus, speex, aac, amr	Any	Unlimited
Fun-ASR-Flash-8K-Realtime series (fun-asr-flash-8k-realtime, etc.)	Binary stream	Same as Fun-ASR-Realtime	8 kHz	Unlimited
qwen3-asr-flash-realtime	Binary stream	pcm, opus	8 kHz, 16 kHz	Unlimited
paraformer-realtime-v2, paraformer-realtime-8k-v2	Binary stream	Same as Fun-ASR-Realtime	paraformer-realtime-v2 any rate; paraformer-realtime-8k-v2 8 kHz	Unlimited

All real-time models accept mono (single-channel) input only.

Offline

Model ID	Input method	Audio format	Sample rate	File size/duration
qwen-audio-3.1-asr-flash-filetrans, qwen-audio-3.0-asr-flash-filetrans	Publicly accessible file URL; 1 URL per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤12 hours (≤2 hours recommended when speaker diarization is enabled)
qwen-audio-3.1-asr-flash	URL / Base64; 1 file per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤5 minutes
qwen-audio-3.0-asr-flash	URL / Base64; 1 file per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤5 minutes
fun-asr	Publicly accessible file URL; 1 URL per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤12 hours (≤2 hours recommended when speaker diarization is enabled)
fun-asr-flash-2026-06-15	URL / Base64; 1 file per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤5 minutes
fun-asr-realtime, fun-asr-realtime-2026-02-28	URL / Base64; 1 file per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	Any	≤2 GB; ≤5 minutes
paraformer-v2, paraformer-8k-v2	Same as Fun-ASR	Same as Fun-ASR	paraformer-v2 /v1 any rate; paraformer-8k-v2 8 kHz only	Same as Fun-ASR
qwen3-asr-flash-filetrans	Publicly accessible file URL; 1 URL per request	aac, amr, avi, flac, flv, m4a, mkv, mov, mp3, mp4, mpeg, ogg, opus, wav, webm, wma, wmv	pcm requires 16 kHz; other formats accept any rate (the server resamples to 16 kHz before recognition)	≤2 GB; ≤12 hours
qwen3-asr-flash	URL / Base64 / local file absolute path; 1 file per request	aac, amr, avi, aiff, flac, flv, mkv, mp3, mpeg, ogg, opus, wav, webm, wma, wmv	pcm requires 16 kHz; other formats accept any rate (the server resamples to 16 kHz before recognition)	≤10 MB; ≤5 minutes