Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Alibaba Cloud / Qwen3.5-LiveTranslate

Qwen3.5-LiveTranslate Flash Realtime

Live audio translation with visual context, 60 supported languages, spoken output in 29 languages and optional voice cloning.

availableClosed weightsSpeech-to-speechSpeech translation
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
29 of 37 specification fields collected
Output formatsPCMSee specification notes below
Languages60 listedFull language list in specifications
Input billingSee ratesAlibaba Model Studio · Singapore · International
Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
2026-05-19Source for Release date (opens in a new tab)

Earliest reviewed regional release; snapshot dates are not assumed to be release dates.

Parameters
Not collected
Architecture
Qwen3.5-Omni Thinker–Talker translation architecture with chunked input and controlled speech-synthesis boundariesSource for Architecture (opens in a new tab)
Languages
View all 60 entries
  • Chinese (zh)
  • English (en)
  • Arabic (ar)
  • German (de)
  • French (fr)
  • Spanish (es)
  • Portuguese (pt)
  • Indonesian (id)
  • Italian (it)
  • Korean (ko)
  • Russian (ru)
  • Thai (th)
  • Vietnamese (vi)
  • Japanese (ja)
  • Turkish (tr)
  • Hindi (hi)
  • Malay (ms)
  • Dutch (nl)
  • Urdu (ur)
  • Norwegian Bokmål (nb)
  • Swedish (sv)
  • Danish (da)
  • Hebrew (he)
  • Finnish (fi)
  • Polish (pl)
  • Icelandic (is)
  • Czech (cs)
  • Filipino (fil)
  • Persian (fa)
  • Cantonese (yue)
  • Greek (el)
  • Afrikaans (af)
  • Asturian (ast)
  • Belarusian (be)
  • Bulgarian (bg)
  • Bengali (bn)
  • Bosnian (bs)
  • Catalan (ca)
  • Cebuano (ceb)
  • Estonian (et)
  • Galician (gl)
  • Gujarati (gu)
  • Croatian (hr)
  • Hungarian (hu)
  • Javanese (jv)
  • Kazakh (kk)
  • Kannada (kn)
  • Kyrgyz (ky)
  • Latvian (lv)
  • Macedonian (mk)
  • Malayalam (ml)
  • Marathi (mr)
  • Punjabi (pa)
  • Romanian (ro)
  • Slovak (sk)
  • Slovenian (sl)
  • Swahili (sw)
  • Tajik (tg)
  • Azerbaijani (az)
  • Ukrainian (uk)
Source for Languages (opens in a new tab)
Context caching
Not collected
API aliases
qwen3.5-livetranslate-flash-realtime, qwen3.5-livetranslate-flash-realtime-2026-05-19Source for API aliases (opens in a new tab)
Data retention
Standard inference prompts and responses may be retained for up to 30 days, with legal/security exceptions. Stateful conversations and files remain until deleted. These translation models are not listed in the reviewed zero-data-retention roster.Source for Data retention (opens in a new tab)
Speech output languages
View all 29 entries
  • Chinese (zh)
  • English (en)
  • Arabic (ar)
  • German (de)
  • French (fr)
  • Spanish (es)
  • Portuguese (pt)
  • Indonesian (id)
  • Italian (it)
  • Korean (ko)
  • Russian (ru)
  • Thai (th)
  • Vietnamese (vi)
  • Japanese (ja)
  • Turkish (tr)
  • Hindi (hi)
  • Malay (ms)
  • Dutch (nl)
  • Urdu (ur)
  • Norwegian Bokmål (nb)
  • Swedish (sv)
  • Danish (da)
  • Hebrew (he)
  • Finnish (fi)
  • Polish (pl)
  • Icelandic (is)
  • Czech (cs)
  • Filipino (fil)
  • Persian (fa)
Source for Speech output languages (opens in a new tab)
Language coverage
60 audio-input/text-output languages; 29 speech-output languagesSource for Language coverage (opens in a new tab)

Current API roster. The launch article enumerates more names than its claimed total of 60.

Stock voice catalog
47 named presetsSource for Stock voice catalog (opens in a new tab)
Request quotas
10 requests/minute; 100,000 tokens/minute in Beijing and SingaporeSource for Request quotas (opens in a new tab)

Published model quotas, not a guarantee of concurrent sessions or actual throughput.

Usage metering
Input audio: 7 tokens/second. Output audio: 12.5 tokens/second. Images: 0.5 token per 32×32 pixel block.Source for Usage metering (opens in a new tab)
Request mode
Realtime API over WebSocket, WebRTC or AOQSource for Request mode (opens in a new tab)
Audio formats
PCM, OpusSource for Audio formats (opens in a new tab)

Input formats. Output uses PCM only.

Audio sample rates
8000, 16000Source for Audio sample rates (opens in a new tab)

Input sample rates; 16 kHz is the default. These are not selectable output rates.

Utterance endpointing
Automatic server VAD by default, or manual audio-buffer commitsSource for Utterance endpointing (opens in a new tab)

VAD threshold range −1 to 1, default 0.2; silence duration 200–6,000 ms, default 1,000 ms. Manual mode uses turn_detection=null.

Source-language transcription
Optional source transcript; qwen3-asr-flash-realtime is enabled by default and can be disabled with nullSource for Source-language transcription (opens in a new tab)

Source-ASR billing for this generation is not explicitly established. The free, always-on transcription statement applies to 3.8.

Translation pipeline
Source language can be supplied or detected automatically; target defaults to EnglishSource for Translation pipeline (opens in a new tab)
Same-language output
Optionally skip text, audio or both when source and target match; only Chinese (zh) or English (en) targetsSource for Same-language output (opens in a new tab)
Custom vocabulary
Source-to-target term mappings; no more than 1,000 hotwords recommendedSource for Custom vocabulary (opens in a new tab)
Image input limits
JPG/JPEG, at most 1080p and 500 KB before Base64 encoding; maximum 2 frames/second; audio must arrive before the first imageSource for Image input limits (opens in a new tab)
Voice cloning
SupportedSource for Voice cloning (opens in a new tab)

Off by default. Live extraction is once per session or before each response; both use voice=default. A previously created voice uses frequency=never.

Voice controls
Tina is the default preset; optionally select another preset, live voice extraction or a previously enrolled voiceSource for Voice controls (opens in a new tab)
Voice enrollment cost
Live extraction has no separate fee. The guide lists USD 0.01 per successfully created advance voice, but the linked enrollment target roster omits LiveTranslate.Source for Voice enrollment cost (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Alibaba Model Studio · Singapore · International

Provider source (opens in a new tab)
API model ID
qwen3.5-livetranslate-flash-realtime
Access
wss://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/api-ws/v1/realtime?model=qwen3.5-livetranslate-flash-realtime. Use a matching regional workspace API key. Realtime translation; WebRTC and AOQ are also documented transports.
Regions
Singapore / International deployment scope
Request quotas
10 RPM; 100,000 TPMSource for Request quotas (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
Realtime translationAudio input1$7.501M tokens
Realtime translationSpeech output1$30.001M tokens
Realtime translationVisual input1$0.551M tokens
Realtime translationTranslated text output1$20.001M tokens

Rate conditions

  1. Audio input 7 tokens/second; speech output 12.5 tokens/second; each 32×32 image block 0.5 token. Translated text and speech have separate listed rates; accompanying-text co-billing is not explicitly resolved. Singapore offers 1M free tokens valid 90 days from the latest of activation, model release or application approval.
Pricing source (opens in a new tab)

Data & enterprise terms

Standard inference prompts and responses may be retained for up to 30 days, with legal/security exceptions. Stateful conversations and files remain until deleted. These translation models are not listed in the reviewed zero-data-retention roster.

Read the service policy (opens in a new tab)

Alibaba Model Studio · Singapore · International

Provider source (opens in a new tab)
API model ID
qwen3.5-livetranslate-flash-realtime-2026-05-19
Access
wss://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/api-ws/v1/realtime?model=qwen3.5-livetranslate-flash-realtime-2026-05-19. Use a matching regional workspace API key. Realtime translation; WebRTC and AOQ are also documented transports.
Regions
Singapore / International deployment scope
Request quotas
10 RPM; 100,000 TPMSource for Request quotas (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
Realtime translationAudio input1$7.501M tokens
Realtime translationSpeech output1$30.001M tokens
Realtime translationVisual input1$0.551M tokens
Realtime translationTranslated text output1$20.001M tokens

Rate conditions

  1. Audio input 7 tokens/second; speech output 12.5 tokens/second; each 32×32 image block 0.5 token. Translated text and speech have separate listed rates; accompanying-text co-billing is not explicitly resolved. Singapore offers 1M free tokens valid 90 days from the latest of activation, model release or application approval.
Pricing source (opens in a new tab)

Data & enterprise terms

Standard inference prompts and responses may be retained for up to 30 days, with legal/security exceptions. Stateful conversations and files remain until deleted. These translation models are not listed in the reviewed zero-data-retention roster.

Read the service policy (opens in a new tab)

Alibaba Model Studio · China (Beijing)

Provider source (opens in a new tab)
API model ID
qwen3.5-livetranslate-flash-realtime
Access
wss://{WorkspaceId}.cn-beijing.maas.aliyuncs.com/api-ws/v1/realtime?model=qwen3.5-livetranslate-flash-realtime. Use a matching regional workspace API key. Realtime translation; WebRTC and AOQ are also documented transports.
Regions
China (Beijing)
Request quotas
10 RPM; 100,000 TPMSource for Request quotas (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
Realtime translationAudio input1$5.5011M tokens
Realtime translationSpeech output1$22.0031M tokens
Realtime translationVisual input1$0.4541M tokens
Realtime translationTranslated text output1$13.7521M tokens

Rate conditions

  1. Audio input 7 tokens/second; speech output 12.5 tokens/second; each 32×32 image block 0.5 token. Translated text and speech have separate listed rates; accompanying-text co-billing is not explicitly resolved. No regional free quota is listed.
Pricing source (opens in a new tab)

Data & enterprise terms

Standard inference prompts and responses may be retained for up to 30 days, with legal/security exceptions. Stateful conversations and files remain until deleted. These translation models are not listed in the reviewed zero-data-retention roster.

Read the service policy (opens in a new tab)

Alibaba Model Studio · China (Beijing)

Provider source (opens in a new tab)
API model ID
qwen3.5-livetranslate-flash-realtime-2026-05-19
Access
wss://{WorkspaceId}.cn-beijing.maas.aliyuncs.com/api-ws/v1/realtime?model=qwen3.5-livetranslate-flash-realtime-2026-05-19. Use a matching regional workspace API key. Realtime translation; WebRTC and AOQ are also documented transports.
Regions
China (Beijing)
Request quotas
10 RPM; 100,000 TPMSource for Request quotas (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
Realtime translationAudio input1$5.5011M tokens
Realtime translationSpeech output1$22.0031M tokens
Realtime translationVisual input1$0.4541M tokens
Realtime translationTranslated text output1$13.7521M tokens

Rate conditions

  1. Audio input 7 tokens/second; speech output 12.5 tokens/second; each 32×32 image block 0.5 token. Translated text and speech have separate listed rates; accompanying-text co-billing is not explicitly resolved. No regional free quota is listed.
Pricing source (opens in a new tab)

Data & enterprise terms

Standard inference prompts and responses may be retained for up to 30 days, with legal/security exceptions. Stateful conversations and files remain until deleted. These translation models are not listed in the reviewed zero-data-retention roster.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Alibaba Model Studio · Singapore · International

qwen3.5-livetranslate-flash-realtime
Not verified

Not established for this translation selector by the reviewed sources.

How it works
Not collected

Not established for this translation selector by the reviewed sources.

Minimum reusable input
Not collected

Not established for this translation selector by the reviewed sources.

Lifetime & refresh
Not collected

Not established for this translation selector by the reviewed sources.

How to enable it
Not collected

Not established for this translation selector by the reviewed sources.

Usage fields
Not collected

Not established for this translation selector by the reviewed sources.

Billing rules
Not collected

Not established for this translation selector by the reviewed sources.

Limits & data boundaries
Not collected

Not established for this translation selector by the reviewed sources.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Alibaba Model Studio · Singapore · International

qwen3.5-livetranslate-flash-realtime-2026-05-19
Not verified

Not established for this translation selector by the reviewed sources.

How it works
Not collected

Not established for this translation selector by the reviewed sources.

Minimum reusable input
Not collected

Not established for this translation selector by the reviewed sources.

Lifetime & refresh
Not collected

Not established for this translation selector by the reviewed sources.

How to enable it
Not collected

Not established for this translation selector by the reviewed sources.

Usage fields
Not collected

Not established for this translation selector by the reviewed sources.

Billing rules
Not collected

Not established for this translation selector by the reviewed sources.

Limits & data boundaries
Not collected

Not established for this translation selector by the reviewed sources.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Alibaba Model Studio · China (Beijing)

qwen3.5-livetranslate-flash-realtime
Not verified

Not established for this translation selector by the reviewed sources.

How it works
Not collected

Not established for this translation selector by the reviewed sources.

Minimum reusable input
Not collected

Not established for this translation selector by the reviewed sources.

Lifetime & refresh
Not collected

Not established for this translation selector by the reviewed sources.

How to enable it
Not collected

Not established for this translation selector by the reviewed sources.

Usage fields
Not collected

Not established for this translation selector by the reviewed sources.

Billing rules
Not collected

Not established for this translation selector by the reviewed sources.

Limits & data boundaries
Not collected

Not established for this translation selector by the reviewed sources.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Alibaba Model Studio · China (Beijing)

qwen3.5-livetranslate-flash-realtime-2026-05-19
Not verified

Not established for this translation selector by the reviewed sources.

How it works
Not collected

Not established for this translation selector by the reviewed sources.

Minimum reusable input
Not collected

Not established for this translation selector by the reviewed sources.

Lifetime & refresh
Not collected

Not established for this translation selector by the reviewed sources.

How to enable it
Not collected

Not established for this translation selector by the reviewed sources.

Usage fields
Not collected

Not established for this translation selector by the reviewed sources.

Billing rules
Not collected

Not established for this translation selector by the reviewed sources.

Limits & data boundaries
Not collected

Not established for this translation selector by the reviewed sources.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

Voice choices

47 voices

Choose a voice within Qwen3.5-LiveTranslate Flash Realtime. These are API voice identifiers, not additional model releases. Source links document each voice.

Published voice catalog
Voice / API IDLanguage & accentDelivery
TinaTinaMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
CindyCindyChinese (Taiwan), English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
Liora MiraLiora MiraMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
SunnybobiSunnybobiMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
RaymondRaymondMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
EthanEthanMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
Theo CalmTheo CalmMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)
SerenaSerenaMandarin Chinese, English, French, German, Russian, Italian, Spanish, Portuguese, Japanese, Korean, Thai, Indonesian, Arabic, Vietnamese, Turkish, Finnish, Polish, Hindi, Dutch, Czech, Urdu, Filipino, Swedish, Danish, Hebrew, Icelandic, Malay, Norwegian, Persian
See the published language list
Provider preset voice. Supported languages vary by voice.Voice source (opens in a new tab)

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Inference: these named translation API releases have their own service contracts and release history, with no corresponding LiveTranslate weights in the reviewed official Qwen inventory. Their use of Omni infrastructure does not identify them with the public Omni checkpoints. Private lineage and parameter counts remain unverified.

Classification evidence (opens in a new tab)

Lifecycle dates

Platform release · Singapore19 May 2026Source (opens in a new tab)
Platform release · China (Beijing)19 May 2026Source (opens in a new tab)

Sources reviewed