Recently, DeepSeek launched its entire series of large models, including DeepSeek-R1, V3, and Coder, on the National Supercomputing Internet Platform (www.scnet.cn). This development is supported by domestic super-intelligent integrated computing power and a national unified computing service system, marking a significant advancement for China's AI large model industry.

The National Supercomputing Internet Platform has introduced various versions of the DeepSeek-R1 model, including 1.5B, 7B, 8B, and 14B parameters. Users can access online inference services easily through QR code scanning or web access, eliminating the need for local deployment. For enterprise-level applications, the platform allows the introduction of proprietary data for customized private model development, with plans to release larger parameter versions like 32B and 70B soon. Additionally, specialized models such as DeepSeek-V3, the Coder series, and the Math series (7B) are designed to address diverse scenarios, such as code generation and mathematical reasoning.

Compared to models developed overseas, the DeepSeek series significantly minimizes computing resource consumption through effective algorithm optimization. When paired with the National Supercomputing Platform's capabilities for affordable and user-friendly computing resource scheduling, this results in a considerable reduction in invocation costs. For instance, the DeepSeek-R1-Distill-Qwen-7B features a WebUI dialogue interface that provides an easy-to-use experience, while enterprise users can seamlessly integrate into production environments via API interfaces, creating a comprehensive link from model invocation to private deployment.

Analysts suggest that the large-scale implementation of DeepSeek models on the supercomputing platform may transform the current landscape of China's AI industry, which has traditionally relied on closed-source models from abroad. As more developers train and optimize localized models based on domestic computational resources, it is anticipated that China will achieve a substantial leap in the application layer of AI technology. This movement, propelled by what some refer to as the "mysterious Eastern power," is introducing new dynamics into the global AI competition.

The incorporation of DeepSeek models into the National Supercomputing Internet Platform not only enhances the accessibility of advanced AI technologies but also fosters a more robust ecosystem for AI development within China. By leveraging local resources and expertise, the potential for innovation and growth in the AI sector is significantly increased.

Key Points

1. DeepSeek has launched multiple AI models on the National Supercomputing Internet Platform.
2. Users can access models easily without local deployment, supporting enterprise customization.
3. The initiative aims to reduce reliance on overseas AI models and strengthen local capabilities.