Changes

Model Studio will begin upgrading the Qwen3-Omni-Flash and Qwen3-Omni-Flash-Realtime models in the Beijing and Singapore regions starting the evening of January 5, 2026 (UTC+08)

After the update, the models will support efficient multimodal understanding of text, images, audio, and video, as well as high-quality speech generation. Key capabilities include:

• Text interaction in 119 languages and speech interaction in 20 languages
• Support for 49 voice styles (newly added in this version)
• Generation of human-like speech for accurate cross-lingual communication
• Strong instruction-following and customizable system prompts, enabling flexible adaptation to dialogue styles and role-specific settings

These enhancements broaden applicability across use cases such as text generation, voice assistants, and multimedia analysis—delivering a natural and seamless multimodal interaction experience.

The specific updates are as follows:

• qwen3-omni-flash will be automatically upgraded to the qwen3-omni-flash-2025-12-01 snapshot model, with no changes to billing items or pricing.
• qwen3-omni-flash-realtime will be automatically upgraded to the qwen3-omni-flash-realtime-2025-12-01 snapshot model, with no changes to billing items or pricing.

Model mapping

If your application is better suited to the current model version, you may manually switch to it before January 5, 2026 (UTC+08).

• qwen3-omni-flash: Current: qwen3-omni-flash-2025-09-15. After update: qwen3-omni-flash-2025-12-01
• qwen3-omni-flash-realtime: Current: qwen3-omni-flash-realtime-2025-09-15. After update: qwen3-omni-flash-realtime-2025-12-01

Log on to Model Studio to try. For more model details, see qwen3-omni-flash，qwen3-omni-flash-realtime.