
Shenzhen Futian and Guangzhou Launch 'AI Civil Servants': Deploying DeepSeek Large Model Covering 240 Scenarios



DeepSeek is raising a second round. Revenue in Jan-July 2026 was ~475M yuan, 10x FY2025; net loss narrowed from 935M to 715M. Open-source models boosted margins and growth, paving a potential IPO. After a 50B yuan first round in June, it launched a second round targeting 50B yuan.....
The internal test interface of Tencent's new generation large model Hy4 has been exposed, with its release imminent. The interface shows that Tencent continues its strategy of self-research and open source, offering three options: the expert-level large model Hy4, the newly upgraded Hy3 suitable for most scenarios, and DeepSeek, which focuses on reasoning and understanding. Hy4 is the focus of this upgrade.
DeepSeek has launched a dedicated WeChat account for its Harness team, signaling the imminent release of Harness products. Harness is a core engineering system enabling large models to operate, offering tool calling, task planning, and execution scheduling, essential for AI agent development.....
Investor Duan Yongping denied investing in large AI models but wished Liang Wenfeng success. Both Zhejiang University alumni, Duan praised Liang, saying 'you can't beat stock trading in the AI era.'....
DeepSeek clarified that the 'secret nicknames' from its Deep Think mode are temporary labels generated during reasoning to aid context understanding and answer optimization, not private user data logging. The issue arose from user-shared screenshots, sparking online debate and trending on social media.....