Qwen3.8-Omni-Flash
Natively multimodal agent model on the Qwen3.8-Next MoE architecture with a 1M-token context, built to plan and finish tasks, not just perceive.
Native multimodal co-training keeps text ability while transferring agentic skill to audio and video. Inherits Qwen3.8-Next's sparse MoE, extends context to 1M tokens, and targets video editing, translation and music-video generation. The report also presents the Qwen-MM-Plugins framework.
- Date
- Thursday 17 September 2026
- Lab
- Alibaba (Qwen)
- Kind
- model
- Access
- closed API
Releasebot and Alibaba Cloud's Model Studio list date the launch to 2026-09-17, with the realtime variant following on 2026-09-21. The Qwen3.8-Omni technical report (arXiv 2609.25611, 'Qwen Team') was submitted 2026-09-22 and confirms the 1M context and the Qwen3.8-Next architecture. Model Studio lists access as a closed API.
Sources
- releasebot.io/updates/qwen
- versely.studio/blog/qwen-4-announced-at-apsara-2026
- alibabacloud.com/help/en/model-studio/newly-released-models
- github.com/QwenLM/Qwen-Live-Harness
- arxiv.org/abs/2609.25611
This record was checked against its sources on 6 October 2026. How we check