Tencent Offers OpenClaw Installation; Xiaomi Introduces Mobile AI Agent; OpenAI Releases ChatGPT 5.4

Tencent hosted a free on-site installation event for OpenClaw, an AI Agent tool, at its Shenzhen headquarters, attracting significant queues. Tencent Cloud Lighthouse engineers provided comprehensive support, including deployment, model configuration, and skill activation, addressing the high barrier to entry for OpenClaw installation. Tencent Cloud Lighthouse lightweight application server is the first cloud platform in China to offer application services for OpenClaw.

Tencent Cloud Lighthouse server racks in a data center, with engineers working.

Qianwen has launched a special promotion for International Women's Day, allowing users to claim an AI service card through the Qianwen App. This initiative mirrors a "3 Billion CNY Treat" approach previously used during the Chinese New Year. The event is scheduled to conclude on March 8.

A hand holding a Xiaomi smartphone displaying the miclaw AI agent interface.

Xiaomi has initiated a small-scale closed beta for Xiaomi miclaw, a mobile-end, system-level intelligent agent. Similar to OpenClaw, Xiaomi miclaw can access mobile system tools, application functionalities, and Xiaomi ecosystem devices with user authorization. It is designed to automatically deconstruct tasks and execute them step-by-step based on user instructions.

Abstract visualization of 3D models rapidly generating, with geometric shapes and light trails.

General AI company VAST announced it has secured $50 million in Series A funding. The company's large model, Smart Mesh, is capable of rapidly generating 5,000-face game models. Its Tripo Studio platform has reportedly attracted over 6.5 million creators and facilitated the generation of nearly 100 million 3D models.

A screen displaying the ChatGPT 5.4 interface with code and text, representing its new capabilities.

OpenAI officially released the ChatGPT 5.4 model, integrating reasoning, programming, and agent workflow capabilities. This marks the first cutting-edge model to natively support computer operations. The model will be progressively rolled out across ChatGPT, API, and Codex platforms. Key enhancements in GPT-5.4 include support for a million-level context window, enabling the processing of extensive codebases or document collections. Portions exceeding the standard 272k context are charged at twice the rate. Additionally, a new "mid-interruption" feature allows users to modify instructions in real-time during model generation.