Alibaba releases Qwen3.8-Omni-Flash, a cheap fast omni model for text, image, audio and video

gaganghotra_ · x · 2026-09-18

Alibaba has released Qwen3.8-Omni-Flash, a cheap and fast omni model that takes text, image, audio and video as input in one brain and answers in text. The company claims roughly a 25% capability improvement over the previous Omni model, and positions it as a Gemini 3.8 Flash-class swap for audio/video understanding.

Related event: Alibaba Qwen Releases Qwen3.8-Omni-Flash, an Agentic Omni-Modal Model(8 posts)→

Original post →

More from Models

Models channel →