LongCat-Flash-Omni officially released and open-sourced: ushering in the era of full-modal real-time interaction.

Meituan's LongCat team officially released and open-sourced the LongCat-Flash-Omni model. Based on the LongCat-Flash series architecture, the model integrates multimodal perception and speech reconstruction modules, boasting a total of 560 billion parameters and 27 billion activation parameters, enabling low-latency real-time audio and video interaction.