Qwen, Alibaba's AI unit, has launched Qwen3.8-Omni-Flash, a next-generation native omnimodal model aimed at strengthening agent capabilities in real-world productivity scenarios. According to the company, the model advances omnimodal systems from merely understanding audio-visual content to planning tasks, calling tools, and completing creative work.
The model is now available on the Qianwen AI Platform and accepts text, image, audio, and video inputs with a 1M-token context window. Qwen says it maintains text performance comparable to a text-only model of the same size while delivering significant gains in omnimodal capabilities.
Across 29 evaluations, Qwen reports that the model's average score improves by more than 25% over Qwen3.5-Omni-Plus. The API price per hour of audio input falls by more than 98%, and the price per hour of audio-visual input drops by more than 93%. For audio-visual agents, coding, and long-horizon tasks, it improves by 36.5 points on WildClawBench-MM and 22.3 points on AgenticVBench, while scoring 69.6 on UniClawBench.
Core capabilities also rise in long-form audio and audio-visual understanding, reasoning, captioning, and multi-speaker recognition. The model gains 8.3 points on LongAudioSpan and 9.6 points on OmniVideoBench, while OmniCap-IF CSR and ISR improve by 8.5 and 14.1 points respectively. AliMeeting DER and cpWER fall from 88.11 / 89.61 to 3.35 / 17.18. Qwen says the model reaches audio-visual performance close to Gemini 3.8 Flash and overall audio performance that exceeds it.
To support these workflows, Qwen expanded Qwen-MM-Plugins with on-demand perception, tool use, and workflow execution for long-form audio and video, and open-sourced Qwen-Live Harness as a native runtime for continuous, real-time omnimodal interaction. Applications span video editing, music video creation, film production and commentary, audio-visual summarization, and real-time conversations.