# LLM
Latest news and articles about LLM
Total: 40 articles found
Application Over Innovation: Chinese AI Models Outpace US Rivals in Global Usage for Three Consecutive Weeks
China has maintained a lead over the United States in global AI model usage for three straight weeks, with Tencent's Hy3 preview now ranking as the most-called model. The data reflects a shift in the AI race toward application and scale, even as domestic competition sees previous leaders like Moonshot AI's Kimi lose ground.
The Utility of Intelligence: Shanghai Telecom Turns AI Tokens into a Commodity
Shanghai Telecom has launched China's first telecom-integrated AI token packages, allowing users to buy 250,000 tokens for one yuan via their mobile bills. This move transforms AI computing power into a standard consumer utility, significantly lowering the barrier for AI adoption in the world's second-largest economy.
The $50 Billion Paradox: DeepSeek’s Reluctant Dive into the Venture Capital Deluge
DeepSeek, the prominent Chinese AI lab founded by Liang Wenfeng, has reportedly abandoned its 'no-funding' policy amid a global investment surge. The company's valuation has surged to an estimated $50 billion in just one month as investors scramble to gain exposure to its industry-leading efficiency and talent.
China’s AI Arms Race Intensifies: StepFun Eyes Hong Kong Debut with $2.5 Billion War Chest
Chinese AI startup StepFun is finalizing a $2.5 billion funding round and dismantling its offshore structure to prepare for a Hong Kong IPO. The move reflects a broader trend of Chinese AI firms seeking strategic industrial backing and local listings amid tightening global tech competition.
Scale Over Spec: Tencent’s Hunyuan Sees Usage Surge as China’s AI War Pivots to Adoption
Tencent's Hunyuan Hy3 preview model has achieved a tenfold increase in token usage within two weeks of release, signaling a successful shift toward large-scale enterprise adoption. This growth underscores an intensifying battle among Chinese tech giants to dominate the AI ecosystem through volume and integration.
Apple's Walled Garden Cracks: iOS 27 to Introduce Choice in AI Models
Apple plans to allow users to select third-party AI models as their primary digital assistants starting with iOS 27. This shift signals a move toward platform neutrality and reflects a strategic response to global regulatory pressures and the need for localized AI solutions.
The End of the Free Lunch: ByteDance’s Doubao Signals a Strategic Shift in China’s AI Wars
ByteDance is transitioning its Doubao AI chatbot to a freemium model by introducing paid subscriptions, marking a strategic shift toward monetization in China's competitive generative AI market.
China Reclaims AI Dominance in Global Usage as Tencent and DeepSeek Surge
Chinese AI models have overtaken US models in global weekly usage volume, driven by massive surges in Tencent's Hy3 and DeepSeek's aggressive price cuts. This shift highlights a strategic pivot toward high-volume adoption and the commoditization of AI tokens in the Chinese tech ecosystem.
The Agent Orchestrator: Tencent’s QClaw Upgrade Bridges the Gap Between Rival Models
Tencent's QClaw upgrade introduces support for the Hermes AI Agent framework and integrates DeepSeek-V4 Pro alongside Tencent's own Hunyuan Hy3. This move highlights a strategic shift toward interoperability and multi-model orchestration in China's intensifying AI market.
Race to the Bottom: DeepSeek Slashing API Costs Signals New Intensity in China's AI Price War
DeepSeek has announced a 90% reduction in its input caching API prices, with additional discounts on its flagship Pro model through 2026. This aggressive pricing strategy signals a new phase in the global AI competition, focusing on cost-efficient scaling for enterprise developers and massive ecosystem adoption.
National Team Play: China’s State-Backed Supercomputing Network Offers Free Access to DeepSeek-V4
China’s National Supercomputing Infrastructure has launched free access to the DeepSeek-V4 model, emphasizing its 1-million-token context capacity. This move integrates state-led computing resources with cutting-edge private AI development to accelerate domestic innovation.
DeepSeek’s V4 Gambit: Low-Cost LLMs and the Pivot to Domestic Powerhouses
DeepSeek has launched its V4 model series, featuring a 1.6 trillion parameter flagship that optimizes million-token contexts at unprecedented price points. The release highlights a deepening partnership with Huawei's Ascend compute platform, signaling a strategic shift toward domestic hardware autonomy amidst global supply constraints.