← All models

Qwen3.8-Omni-Flash

Published Publisher Alibabaqwen3.8-omni-flash

Overview

Qwen3.8-Omni-Flash is a native multimodal model for audio and video understanding, content analysis, and agent applications. It accepts text, images, audio, and video and produces text. Alibaba Cloud Model Studio provides its inference service. (alibabacloud.com)

Capabilities and uses

  • Supports thinking and non-thinking modes, function calling, web search, and context caching. Thinking is enabled by default.
  • Supports spatial multichannel audio input and audio input in 113 languages and dialects.
  • Uses include audio/video analysis, question answering, meeting summaries, subtitles, and multimodal agent workflows. It is available through the Chat Completions and Responses APIs. (alibabacloud.com)

Limits

  • The model outputs text only; it does not directly generate audio. The separately identified qwen3.8-omni-flash-realtime is intended for real-time audio/video conversation.
  • Context window: 1M tokens. Maximum input: 991,808 tokens in non-thinking mode or 983,616 tokens in thinking mode. Maximum output: 131,072 tokens.
  • Alibaba's current guide lists audio up to 3 hours and video up to 2 hours for content analysis; the original launch announcement described individual audio or video inputs as approximately 1 hour maximum. Check current API requirements and rate limits for the selected region. (help.aliyun.com)

Release date

September 17, 2026, according to Alibaba Cloud's launch announcement and model update listing. (aliyun.com)

Official pricing

Prices below are for qwen3.8-omni-flash, per 1 million tokens; cache-hit prices apply to eligible cached input. The domestic and international sites are separate offerings, not currency conversions. Checked November 16, 2026.

Official site and region Deployment scope Input Cache-hit input Output
China site, Beijing Chinese mainland CNY 0.8 CNY 0.1 CNY 2.7
China site, Singapore International CNY 1.094 CNY 0.117 CNY 3.427
China site, Hong Kong, Tokyo, Frankfurt or Virginia Global CNY 0.8 CNY 0.1 CNY 2.7
International site, Singapore International USD 0.15 USD 0.016 USD 0.47
International site, Hong Kong, Tokyo, Frankfurt or Virginia Global USD 0.113 USD 0.014 USD 0.382

Different modalities have their own token calculation rules. The listed cache-hit rate is not the standard input rate. (help.aliyun.com)

Official sources