Skip to content
TrustList
QA
AI Model

Qwen-Audio-3.1-Realtime-Plus

New· 4

Alibaba's proprietary full-duplex speech model on Model Studio, with function calling, web search and voice cloning

About Qwen-Audio-3.1-Realtime-Plus

Voice assistants usually wait for a person to finish speaking before they reply. Qwen-Audio-3.1-Realtime-Plus, from Alibaba, is a full-duplex speech-to-speech model that listens and speaks at the same time.

How the model connects

The model is end to end and is served through Alibaba Cloud Model Studio. Audio streams in, and speech and text stream back, over one persistent WebSocket connection; Alibaba's own AOQ protocol and WebRTC are supported too. Input must be 16 kHz mono PCM, and the output is 24 kHz PCM. Model Studio's lifecycle page gives 20 September 2026 as the release date, for the Singapore (international) and China (Beijing) regions alike, with a public announcement following on 23 September 2026 under the Qwen Audio 3.1 banner.

Features and conversation limits

Alibaba lists a 262,144-token context window along with function calling, web search and voice cloning. Web search cannot be enabled at the same time as function calling. Turn detection can be acoustic, semantic or push-to-talk. Eleven languages are supported, among them English, German, French, Spanish, Japanese and Korean, along with 21 Chinese varieties, Mandarin and Cantonese included. The model retains conversation history for at most 50 turns (the default is 20) or 300 seconds of audio in total.

Upgrading and intended uses

The integration protocol is unchanged from Qwen-Audio-3.0-Realtime-Plus, so existing integrations need only a new model name, and eight more system voices are added. Alibaba positions the model for voice assistants, lines run by customer service, and AI companions.

Billing and access

Pricing is per million tokens and differs for text and audio, input and output, with mainland China rates lower than Singapore. Audio converts at 12.5 tokens per second, and since earlier turns are billed again as input with every new turn, long sessions grow steadily dearer. A free quota applies in Singapore only. The model is proprietary, with no downloadable weights and no published parameter count.

TrustList summary

Qwen-Audio-3.1-Realtime-Plus is Alibaba's speech-to-speech model that handles audio in both directions at once, served through Alibaba Cloud Model Studio. Its lifecycle listing dates the release to 20 September 2026 in both the Singapore and China (Beijing) regions, and Alibaba announced it publicly on 23 September.

Features

  • Streaming voice: receives audio as a stream and sends back speech and text as a stream over a persistent WebSocket link, and also supports AOQ and WebRTC.
  • Tools and search: function calling and web search, although web search cannot run at the same time as function calling.
  • Voice cloning: clones voices, and adds eight system voices to the existing set.
  • Turn detection: acoustic, semantic (smart turn) or push-to-talk.
  • Conversation history: retained for up to 50 turns (20 by default), or until 300 seconds of audio have accumulated.

Who it works with

It is intended for voice assistants, customer-service lines and AI companions.

Key facts

  • Context window: 262,144 tokens.
  • Languages: eleven languages, including English and Japanese, and 21 Chinese varieties including Mandarin and Cantonese.
  • Audio format: 16 kHz mono PCM in, and 24 kHz PCM out.
  • Weights: proprietary, with no downloadable weights.

Summary written by TrustList from the business's own published information, Nov 2026.

Benchmarks & AI stats

BenchmarkOfficialCommunity avg
No benchmark scores yet. Be the first to add one.

“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).

Sign in to add a benchmark score for this model.

Qwen-Audio-3.1-Realtime-Plus launches

Trust Score

4/ 100
Trust Score: New

An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.

Verification
0/100 · 20%
Reviews
0/100 · 30%
Awards
0/100 · 15%
Transparency
18/100 · 20%
Engagement
0/100 · 15%
Joining soon
Recommendations
Coming soon
Complaints
Coming soon

Updated 9/25/2026

Request a demo or quote from Qwen-Audio-3.1-Realtime-Plus

Protected by reCAPTCHA - Google Privacy Policy and Terms apply.

By sending, you agree we may share your request and contact details with the provider once you confirm your email.

Reviews

Write the first review of Qwen-Audio-3.1-Realtime-Plus

Used it? Your experience helps other buyers decide.

Write a review

Questions & answers

No questions yet. Be the first to ask about Qwen-Audio-3.1-Realtime-Plus.

Qwen-Audio-3.1-Realtime-Plus alternatives

Similar listings buyers compare against this one.

Open-weight translation model from Bilibili's Index team, covering 150 languages and following glossary and style instructions

Open-source 2B decision model from AWS's Strands Agents project that picks options and returns scores with a reliability figure

Anthropic's smallest and fastest Claude 5.5 model, available on its own API and three cloud platforms

Cantina Security's first open-weight security model, a post-trained GLM-5.3-Flash for vulnerability research and exploit checks

Black Forest Labs' image generation and editing model with bounding-box layouts, multiple reference images and 4K output

Microsoft AI's real-time speech-to-text model for voice agents, transcribing 60 languages with continuous language detection

Liquid AI's open-weight decision model, giving yes or no, choice or score answers from text and images

Microsoft AI's text-to-speech model for keeping one voice across 23 languages, with voice cloning from short audio