Group 1 - Agentic Search launched by Mita Technology allows users to perform 5-15 tool calls autonomously in one search, with a model parameter scale increased by 10 times compared to the previous version [2] - DingTalk AI Table Assistant officially launched, enabling users to create AI tables and supporting workflows from platforms like Bailian and Coze [2] - Genie Operator-1 (GO-1) by Zhiyuan Robot is now fully open-sourced, marking the first global release of a general-purpose embodied intelligence model using the ViLLA architecture [2] - Alibaba's Qianwen has updated three large models, including the open-source multimodal model Qwen3-Omni, which shows significant advantages in audio tasks compared to similar models [2]