# Large%20Language%20Models
Latest news and articles about Large%20Language%20Models
Total: 89 articles found
China’s Zhipu AI Escalates Open-Source Coding Wars with GLM-5.2 Release
Zhipu AI has announced the release of GLM-5.2 to its entire coding-focused user base, with plans for an API launch and an open-source release under the MIT license next week. This move highlights the firm's strategy to dominate the developer ecosystem through permissive licensing and rapid model iteration.
Huawei’s AI Gambit: HarmonyOS 7 Signals the Dawn of the ‘Agent’ Era
Huawei has launched HarmonyOS 7, billed as the first OS fully rebuilt for the AI era, featuring a new 'Agent' architecture and advanced LLM integration. Following the success of HarmonyOS 6 as China's second-largest mobile OS, this release underscores Huawei’s strategy to lead the global shift toward autonomous, AI-driven operating systems.
From Pixels to Pathologies: Why China’s Tech Giants are Re-Engaging the Medical Frontier
China's tech giants are pivoting from shallow digital healthcare services to deep AI integration and physical infrastructure, led by ByteDance's $825 million hospital investment. This shift reflects a move to utilize Large Language Models to solve chronic medical resource shortages and capture a rapidly growing market.
Silicon Sovereignty: Shenzhen’s Trillion-Parameter Breakthrough Signals New Era for Chinese AI
Researchers in Shenzhen have successfully trained the 1.6-trillion-parameter DeepSeek-V4-Pro model using Huawei’s Ascend 910C chips. This milestone proves that China’s domestic computing clusters are now capable of handling top-tier AI training tasks, marking a significant step in the country's drive to overcome Western semiconductor restrictions.
Hong Kong Unveils 'Super Agent' AI in Bid to Secure High-Tech Productivity Hub Status
Hong Kong has launched HKGAI V3 and its first productivity-grade 'super agent,' a government-funded AI system capable of autonomous operation for 28 hours. The model represents a strategic shift toward integrated, high-stability AI tools for professional reasoning and software development.
Domestic Ambitions: Zhipu AI Signals Strategic Shift with Shanghai STAR Market IPO Plan
Chinese AI heavyweight Zhipu AI has proposed a major IPO on Shanghai's STAR Market, aiming to issue up to 8% of its equity to raise an estimated 15 billion RMB. This move signals a broader trend of Chinese AI unicorns seeking domestic listings to secure the capital needed for the resource-heavy development of large language models.
The Machine's Fingerprint: Why AI Storytelling Remains Trapped in a Narrative Cul-de-Sac
A major study by the University of Maryland and Google DeepMind has identified deep-seated 'narrative fingerprints' in AI writing that cannot be hidden by stylistic prompts. Researchers found that AI models follow rigid, preachy, and linear structures that fundamentally differ from the complex, non-linear, and ambiguous nature of human storytelling.
The Race to the Bottom: Xiaomi Slashes AI Prices by 99% as China’s Token War Escalates
Xiaomi has permanently cut its AI model API prices by up to 99%, joining a brutal price war among Chinese tech giants. This move aims to democratize AI access while leveraging Xiaomi's strong financial position to dominate the hardware-integrated AI ecosystem.
Beyond the Cloud: Why China’s Leading AI Scholars are Abandoning the 'OpenAI Clone' Strategy
Tsinghua professor Liu Zhiyuan argues that China must move beyond cloning OpenAI by focusing on 'intelligence density' and edge computing. He emphasizes that China's competitive advantage lies in integrating high-efficiency small models into real-world industrial scenes rather than competing solely on cloud-based scale.
Hardware as the AI Horizon: China’s ‘Little Dragon’ Prepares for a $10 Billion Hong Kong Debut
Chinese AI startup StepFun is finalizing a massive funding round and preparing for a $10 billion Hong Kong IPO, supported by a coalition of hardware giants and strategic investors aiming to integrate LLMs into physical devices.
The Token Economy: How China’s Inland Frontiers are Mining the New Currency of AI
China's inland cities are bypassing traditional industrial development by leveraging the AI 'token' as a new economic engine. By providing the low-cost computing power and human-in-the-loop services necessary for large language models, these regions are redefining regional growth in the age of generative AI.
Racing the Clock: Zhipu AI Launches High-Speed GLM-5.1 to Capture the Real-Time Enterprise Market
Zhipu AI has launched GLM-5.1-highspeed, an enterprise-grade API delivering 400 tokens per second. The model targets latency-critical applications like AI programming and real-time voice, positioning Zhipu to lead the efficiency-focused second wave of China's AI evolution.