Speech-to-text models

Migrate from closed-source models

Replacing Whisper, Deepgram, or Google speech recognition? Use these QwenCloud equivalents.

Use case	Closed-source examples	QwenCloud recommendation
Real-time recognition	Deepgram Nova-3, Google Chirp 3	`qwen-audio-3.0-asr-flash-streaming`
Offline / file transcription	OpenAI gpt-4o-transcribe, Whisper	`qwen-audio-3.0-asr-flash-filetrans`, `qwen-audio-3.0-asr-flash`

This page walks you through model selection: start with "Decision dimensions" (real-time/offline, domain terminology, speaker diarization, emotion recognition) to identify your scenario; check "Recommended models" for top picks; expand "All models" for full version lists; and verify input constraints in "Audio specifications". Supported languages (including dialects) are listed within each model family under "All models".

Decision dimensions

Use these 4 dimensions to narrow your choice. Each recommends the best-fit model.

Real-time or offline?

Real-time means outputting recognition results while the user is still speaking. Offline means transcribing after the recording ends.

Real-time (streaming recognition) — Uses a WebSocket connection to stream audio in and text out. Ideal for live captions, voice assistants, and meeting transcription. Recommended model: qwen-audio-3.0-asr-flash-streaming (hot words, prompt context, multilingual with dialects).
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-3.0-asr-flash-filetrans (hot words, prompt context, speaker diarization).

Qwen-Audio-3.0-ASR-Flash-Streaming, Qwen-Audio-3.0-ASR-Flash-Filetrans, and the Fun-ASR and Qwen-ASR real-time models support the DashScope SDK (Java, Python). All other models require direct WebSocket or HTTP API calls. Qwen-Audio-3.0-ASR-Flash-Streaming and Fun-ASR-Realtime series models also support the AOQ protocol. For client-side integration that prioritizes stable latency, resilience on weak networks, and built-in full-duplex noise suppression and echo cancellation, AOQ is recommended. For a protocol comparison, see Realtime API overview. For real-time integration, see Real-time speech recognition. For offline integration, see Audio file transcription.

Handling domain terminology

Two approaches, ranked by flexibility:

Prompt context injection — Describe your domain in the system prompt. No setup required -- the model adapts per request. Recommended: the Qwen-Audio-3.0-ASR-Flash-Streaming, Qwen-Audio-3.0-ASR-Flash-Filetrans, and Qwen-Audio-3.0-ASR-Flash series.
Hot words — Supply a weighted vocabulary list. Recommended: the Qwen-Audio-3.0-ASR-Flash-Streaming, Qwen-Audio-3.0-ASR-Flash-Filetrans, and Qwen-Audio-3.0-ASR-Flash series.

Speaker diarization

The Qwen-Audio-3.0-ASR-Flash-Filetrans series (qwen-audio-3.0-asr-flash-filetrans) and Fun-ASR offline models (fun-asr, fun-asr-mtl) support speaker diarization. To identify who said what, use qwen-audio-3.0-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

The top picks for each scenario. Visit the model gallery for details.

Model	Mode	API	Accuracy boost	Emotion	Speaker diarization	Languages	Max duration/size
`qwen-audio-3.0-asr-flash-streaming`	Real-time	WebSocket	Hot words, prompt context	✗	✗	Multilingual with dialects	Unlimited
`qwen-audio-3.0-asr-flash-filetrans`	Offline	HTTP	Hot words, prompt context	✗	✓	Multilingual with dialects	12h / 2GB

All models

Qwen-Audio-3.0-ASR-Flash-Streaming

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

Supported languages: 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-3.0-ASR-Flash-Filetrans

Model	Mode	API	Accuracy boost	Emotion	Speaker diarization	Languages	Max duration/size
`qwen-audio-3.0-asr-flash-filetrans`	Offline	HTTP	Hot words, prompt context	✗	✓	Multilingual with dialects	12h / 2GB

Qwen-Audio-3.0-ASR-Flash

Model	Mode	API	Accuracy boost	Emotion	Speaker diarization	Languages	Max duration/size
`qwen-audio-3.0-asr-flash`	Offline	HTTP	Hot words, prompt context	✗	✗	Multilingual with dialects	5 min / 2GB

Fun-ASR

Model	Mode	API	Accuracy boost	Emotion	Speaker diarization	Languages	Max duration/size
`fun-asr-realtime`	Real-time	WebSocket	Hot words	✗	✗	Multilingual with dialects	Unlimited
`fun-asr-realtime-2026-02-28`	Real-time	WebSocket	Hot words	✗	✗	Chinese, English, Japanese	Unlimited
`fun-asr-realtime-2025-11-07`	Real-time	WebSocket	Hot words	✗	✗	Multilingual with dialects	Unlimited
`fun-asr-realtime-2025-09-15`	Real-time	WebSocket	Hot words	✗	✗	Chinese, English	Unlimited
`fun-asr-flash-8k-realtime`	Real-time	WebSocket	Hot words	✗	✗	Chinese	Unlimited
`fun-asr-flash-8k-realtime-2026-01-28`	Real-time	WebSocket	Hot words	✗	✗	Chinese	Unlimited
`fun-asr`	Offline	HTTP	Hot words	✗	✓	Multilingual with dialects	12h / 2GB
`fun-asr-2025-11-07`	Offline	HTTP	Hot words	✗	✓	Multilingual with dialects	12h / 2GB
`fun-asr-2025-08-25`	Offline	HTTP	Hot words	✗	✓	Chinese, English	12h / 2GB
`fun-asr-mtl`	Offline	HTTP	Hot words	✗	✓	Multilingual with dialects	12h / 2GB
`fun-asr-mtl-2025-08-25`	Offline	HTTP	Hot words	✗	✓	Multilingual with dialects	12h / 2GB
`fun-asr-flash-2026-06-15`	Offline	HTTP	Prompt context	✗	✗	Multilingual with dialects	5 min / 2GB

Supported languages (by version):

Fun-ASR-Realtime main versions (fun-asr-realtime, fun-asr-realtime-2026-02-28, fun-asr-realtime-2025-11-07): 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-realtime-2025-09-15: Chinese (Mandarin), English
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, fun-asr-2025-11-07): 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
fun-asr-2025-08-25: Chinese (Mandarin), English
Fun-ASR(MTL) (fun-asr-mtl, fun-asr-mtl-2025-08-25): Chinese (Mandarin, Cantonese), 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	Mode	API	Accuracy boost	Emotion	Speaker diarization	Languages	Max duration/size
`qwen3-asr-flash-realtime`	Real-time	WebSocket	✗	✓	✗	Multilingual with dialects	Unlimited
`qwen3-asr-flash-realtime-2026-02-10`	Real-time	WebSocket	✗	✓	✗	Multilingual with dialects	Unlimited
`qwen3-asr-flash-realtime-2025-10-27`	Real-time	WebSocket	✗	✓	✗	Multilingual with dialects	Unlimited
`qwen3-asr-flash-filetrans`	Offline	HTTP	✗	✓	✗	Multilingual with dialects	12h / 2GB
`qwen3-asr-flash-filetrans-2025-11-17`	Offline	HTTP	✗	✓	✗	Multilingual with dialects	12h / 2GB
`qwen3-asr-flash`	Offline	HTTP (OpenAI compatible)	✗	✓	✗	Multilingual with dialects	5 min / 10MB
`qwen3-asr-flash-2026-02-10`	Offline	HTTP (OpenAI compatible)	✗	✓	✗	Multilingual with dialects	5 min / 10MB
`qwen3-asr-flash-2025-09-08`	Offline	HTTP (OpenAI compatible)	✗	✓	✗	Multilingual with dialects	5 min / 10MB

Supported languages: All Qwen-ASR models (qwen3-asr-flash-realtime, qwen3-asr-flash-filetrans, qwen3-asr-flash and their snapshot versions) support 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 earlier-generation ASR model family. If your workload allows, migrate to the recommended Fun-ASR or Qwen-ASR models above.

Model	API	Description
`paraformer-realtime-v2`	WebSocket	Real-time recognition; Chinese, English, Japanese, Korean, German, French, Russian
`paraformer-realtime-v1`	WebSocket	Real-time recognition; Chinese, English, Japanese, Korean, German, French, Russian
`paraformer-realtime-8k-v2`	WebSocket	Real-time recognition; 8 kHz telephony; Chinese
`paraformer-realtime-8k-v1`	WebSocket	Real-time recognition; 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; 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-v1, paraformer-realtime-8k-v2, paraformer-realtime-8k-v1, paraformer-8k-v2: Chinese (Mandarin)

Audio specifications

The tables below summarize audio specs (input method, format, sample rate, size/duration) for real-time and offline modes. For supported languages, see the model family sections under "All models" above.

Real-time

Model	Input method	Audio format	Sample rate	Size/duration
Qwen-Audio-3.0-ASR-Flash-Streaming (`qwen-audio-3.0-asr-flash-streaming`)	Binary stream	`pcm`, `wav`, `mp3`, `opus`, `speex`, `aac`, `amr`	Any	Unlimited
Fun-ASR-Realtime (`fun-asr-realtime` family)	Binary stream	`pcm`, `wav`, `mp3`, `opus`, `speex`, `aac`, `amr`	Any	Unlimited
Fun-ASR-Flash-8K-Realtime (`fun-asr-flash-8k-realtime` family)	Binary stream	Same as Fun-ASR-Realtime	8 kHz	Unlimited
Qwen-ASR-Realtime (`qwen3-asr-flash-realtime` family)	Binary stream	`pcm`, `opus`	8 kHz, 16 kHz	Unlimited
Paraformer-Realtime (`paraformer-realtime-v2/v1`, `paraformer-realtime-8k-v2/v1`)	Binary stream	Same as Fun-ASR-Realtime	`paraformer-realtime-v2` any; `paraformer-realtime-v1` 16 kHz; `paraformer-realtime-8k-*` 8 kHz	Unlimited

All real-time models require mono input.

Offline

Model	Input method	Audio format	Sample rate	File size/duration
Qwen-Audio-3.0-ASR-Flash-Filetrans (`qwen-audio-3.0-asr-flash-filetrans`)	Publicly accessible file URL, 1 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 with speaker diarization)
Qwen-Audio-3.0-ASR-Flash (`qwen-audio-3.0-asr-flash`)	URL / Base64, 1 per request	`aac`, `amr`, `avi`, `flac`, `flv`, `m4a`, `mkv`, `mov`, `mp3`, `mp4`, `mpeg`, `ogg`, `opus`, `wav`, `webm`, `wma`, `wmv`	Any	≤2 GB; ≤5 min
Fun-ASR (`fun-asr`, `fun-asr-mtl` family)	Publicly accessible file URL, 1 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 with speaker diarization)
Fun-ASR-Flash (`fun-asr-flash-2026-06-15`)	URL / Base64, 1 per request	`aac`, `amr`, `avi`, `flac`, `flv`, `m4a`, `mkv`, `mov`, `mp3`, `mp4`, `mpeg`, `ogg`, `opus`, `wav`, `webm`, `wma`, `wmv`	Any	≤2 GB; ≤5 min
Paraformer (`paraformer-v2/v1`, `paraformer-mtl-v1`, `paraformer-8k-v2/v1`)	Same as Fun-ASR	Same as Fun-ASR	`paraformer-v2` any; `paraformer-8k-*` 8 kHz only	Same as Fun-ASR
Qwen3-ASR-Flash-Filetrans (`qwen3-asr-flash-filetrans` family)	Publicly accessible file URL, 1 per request	`aac`, `amr`, `avi`, `flac`, `flv`, `m4a`, `mkv`, `mov`, `mp3`, `mp4`, `mpeg`, `ogg`, `opus`, `wav`, `webm`, `wma`, `wmv`	`pcm` must be 16 kHz; other formats any (server resamples to 16 kHz)	≤2 GB; ≤12 hours
Qwen3-ASR-Flash (`qwen3-asr-flash` family)	URL / Base64 / local file absolute path, 1 per request	`aac`, `amr`, `avi`, `aiff`, `flac`, `flv`, `mkv`, `mp3`, `mpeg`, `ogg`, `opus`, `wav`, `webm`, `wma`, `wmv`	`pcm` must be 16 kHz; other formats any (server resamples to 16 kHz)	≤10 MB; ≤5 min

Need speech translated to a different language? See Speech-to-Speech models for real-time and file-based translation with LiveTranslate and Qwen-Omni.

Learn more

Real-time ASR

Stream audio and get text back in real time.

Audio file transcription

Transcribe recorded audio files with async APIs.

Improve accuracy

Boost recognition of domain-specific terms.

​Migrate from closed-source models

​Decision dimensions

​Real-time or offline?

​Handling domain terminology

​Speaker diarization

​Emotion recognition

​Recommended models

​All models

​Audio specifications

​Real-time

​Offline

​Learn more

Real-time ASR

Audio file transcription

Improve accuracy

Migrate from closed-source models

Decision dimensions

Real-time or offline?

Handling domain terminology

Speaker diarization

Emotion recognition

Recommended models

All models

Audio specifications

Real-time

Offline

Learn more