Back to News
ProductAI Understanding briefing

Alibaba Qwen Team Releases Qwen3.8-LiveTranslate for Real-Time Interpretation

Alibaba's Qwen Team has released Qwen3.8-LiveTranslate, a real-time simultaneous interpretation model that reduces average lag to 2.3 seconds across 60 languages.

4 min readRead the linked source
Source-provided image accompanying Alibaba Qwen Team Releases Qwen3.8-LiveTranslate for Real-Time Interpretation
Source referenceSource recorded
Publisher
marktechpost.com
Source link
marktechpost.comhttps://www.marktechpost.com/2026/09/19/alibaba-qwen-team-releases-qwen3-8-livetranslate/amp/
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Start here

Key terms

API (Application Programming Interface)
A structured way for one software system to send requests to and receive responses from another system.
Context Window
The maximum amount of input tokens a language model can process at once.
Feature
An input variable used by a model to make predictions.
Test yourselfWhat is AI? Quiz

What happened

Alibaba's Qwen Team released Qwen3.8-LiveTranslate, a real-time interpretation model that listens to live speech and returns translated text and speech while the speaker is still talking. The model features a new Interleave architecture, which improves faithfulness, fluency, and conciseness, reducing average lag from 2.8 seconds to 2.3 seconds. It is available as a hosted API on Alibaba Cloud Model Studio and QwenCloud.

Qwen3.8-LiveTranslate introduces an Interleave architecture that processes live speech (and optional video frames) to produce translated text and audio in real time. It supports 60 languages for text output and 29 languages for both text and audio output, handling speaker diarization, synchronized bilingual display, and long‑context disambiguation.

The model is deployed as a hosted API on Alibaba Cloud Model Studio and QwenCloud under the identifier qwen3.8-livetranslate-flash-realtime, accessible via a WebSocket Realtime API. Default audio settings are 16 kHz PCM input and 24 kHz PCM output, with a default voice named Tina.

Pricing varies by region; Beijing rates are listed in USD (e.g., $5.653, $0.466, $14.133, $22.613). Audio input consumes 7 tokens per second, audio output 12.5 tokens per second, making an hour of bidirectional speech cost roughly $1.54 in Singapore before text and image token charges. The is 53,248 tokens (49,152 input, 4,096 output).

Source details: marktechpost.com ↗

Why it matters

Qwen3.8-LiveTranslate is significant because it offers real-time simultaneous interpretation across 60 languages, with measurable improvements in latency, faithfulness, fluency, and conciseness. This capability can transform international communications, live events, and global business meetings by delivering near-instant translation, reducing barriers to multilingual interaction and enabling more seamless cross‑language collaboration.

The 18% reduction in average lag (from 2.8 s to 2.3 s) makes the model more suitable for scenarios where immediacy is critical, such as live conferences, diplomatic briefings, and emergency response communications.

Support for a broad set of languages—including major world languages for both text and audio—expands the model's applicability across diverse markets and user groups, potentially lowering the cost and complexity of multilingual services.

Deployment as a hosted API simplifies integration for developers, allowing rapid adoption without the need for on‑premise infrastructure, while transparent token‑based pricing provides clearer cost forecasting for enterprises.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Document Size:128K tokens
Needle Placement Depth (Location in document):50% into text
Attention Context Buffer Map:
Target Fact (50%)
Equivalent Pages~320Standard book pages
Retrieval Accuracy99.9%Needle recall score
RAM / KV Cache5.1 GBMemory overhead
Prompt CachingActive~80% discount on reuse
Core takeaway: Million-token context windows allow querying whole codebases or legal archives in one prompt. However, KV cache memory scales with context length, making prompt caching crucial for real-time production.
Interactive Concept Check+10 Points
What is AI? Quiz

As use of AI scales up across an organization, what tends to matter most?

What to watch next

Future adoption of Qwen3.8-LiveTranslate in industries that rely on live multilingual communication, further enhancements to language coverage and latency, and competitive responses from other real‑time translation providers.

Adoption rates in sectors like travel, education, and multinational enterprises, where real‑time translation can drive efficiency and user experience improvements.

Potential updates that increase language coverage, further reduce latency, or add offline capabilities, which could broaden the model's use cases.

Competitive dynamics as other AI firms release or improve their own real‑time translation solutions, influencing market pricing and sets.

Related guides & quizzes

What is AI?AI Models ExplainedAI TranslationTest what you know — try a free AI quizLook up an AI term in our glossary
Found this useful?