IT Morning Report 0816: DeepSeek V4 Pro Official Version, Harness Launch on National Supercomputing Internet; Tomato Novel Author 'Resurrects' Three Years After 'Death'; Reportedly Xiaomi 18 Phone Standard Version Delayed to Early Next Year...

The National Supercomputing Internet platform has officially integrated the DeepSeek V4 Pro large model and the Harness Agent Framework. This move marks that domestic large models have achieved full-process coverage from training to deployment on national-level computing infrastructure, providing a standardized technical foundation for research institutions and commercial applications.

This update focuses on strengthening Agent capabilities, meaning the model has seen significant improvements in multi-step task planning and autonomous execution. Performance benchmarked against international advanced models indicates that domestic large models are closing the gap in technical iteration speed, helping to reduce domestic developers' reliance on overseas models.

Relying on the computing resources of the National Supercomputing Internet, the deployment threshold for DeepSeek V4 Pro will be further lowered. This will accelerate the application and implementation of AI technology in vertical industries, especially in research computing and complex business scenarios requiring high computing power support, where the competitiveness of domestic solutions is expected to increase.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.