Tencent Plans Larger Hy4 Model After Hy3 Usage Jumps 68-Fold
The AMW Read
Incremental update to Tencent's model lineup, but signals strategic push in foundation models with ecosystem integration.
Tencent Plans Larger Hy4 Model After Hy3 Usage Jumps 68-Fold
Tencent announced that weekly usage of its Hy3 model has surged more than 68 times compared to its predecessor, following the model's transition from preview to a formal release. The company confirmed plans to launch a larger-parameter Hy4 model in the near term, though no release date or technical specifications were disclosed. Hy3 is now integrated into Tencent products including WorkBuddy, CodeBuddy, Yuanbao, and ima, signaling broad internal deployment across its enterprise and consumer AI offerings.
The sharp usage increase and rapid product integration position Tencent as a significant player in China's foundation model race. By embedding Hy3 across productivity and coding tools, Tencent is creating a captive distribution channel that can drive adoption and feedback loops, similar to how other hyperscalers leverage their ecosystems. The planned Hy4 suggests Tencent is betting on scale and continuous iteration to compete with other domestic labs like Alibaba's Qwen and DeepSeek, even as it expands its model's reach into enterprise workflows.
For builders and investors, Tencent's move underscores the strategic value of distribution in AI. Developers should watch for Hy4's API availability and whether it offers competitive pricing or performance, as Tencent's scale could pressure smaller model providers. Enterprises already using Tencent's suite may see improved AI capabilities integrated into their existing tools, potentially reducing the need for separate AI vendors. Investors should monitor how Tencent's compute investments and model roadmap align with the broader trend of Chinese hyperscalers consolidating AI offerings.