qwen3-asr-flash-realtime

Updated at:

The realtime version of Qwen3-ASR-Flash, a multilingual speech recognition model built on large language models with high accuracy, intelligence, and robustness. Leveraging a powerful base model, massive text and multimodal data, and tens of millions of hours of audio data, Qwen3-ASR-Flash achieves high-precision speech recognition, automatically detecting languages and accurately recognizing speech in multiple languages, with accurate transcription even in complex acoustic environments.

Inference Service Provider

The inference service provider for qwen3-asr-flash-realtime is Alibaba Cloud Model Studio.

Model Capabilities

CapabilitySupportCapabilitySupport

Input Modality

Audio

Output Modality

Text

Model Experience

Unsupported

Function Calling

Unsupported

Structured Outputs

Unsupported

Web Search

Unsupported

Prefix Completion

Unsupported

Context Caching

Unsupported

Batch Inference

Unsupported

Fine-tuning

Unsupported

Context Limits

ParameterValueParameterValue

Max Input Length

—

Max Output Length

—

Context Window

—

Pricing

This page only shows the original pricing for model API calls, excluding any limited-time promotions. Visit Model Studio Console for promotional offers.

China (Beijing)

Billing ItemPrice (CNY)Unit

Audio Duration

0.00033

Per second

Singapore

Scope: International

Billing ItemPrice (CNY)Unit

Audio Duration

0.00066

Per second

How to call the model

Rate Limits

China (Beijing)

ParameterValue

RPM (Requests Per Minute)

1200

Singapore

Scope: International

ParameterValue

RPM (Requests Per Minute)

1200