DeepSeek

DeepSeek unexpectedly updated its R1 model, rather than releasing R2. This trial upgrade, accessible via their website and apps, maintains the existing API. Early tests of the open-source R1 (0528) version show significant improvements, especially in programming, now rivaling international standards. The updated model excels in varied tasks from web design to text processing, sometimes outperforming other models like Claude 4.

DeepSeek, a Beijing-based AGI startup, announced plans to open-source five production-grade code repositories, including training frameworks and inference engines, positioning itself as a transparent alternative to Silicon Valley secrecy. Despite its rapid growth—offering GPT-4-level tools freely while monetizing enterprise APIs—the company faces scrutiny over data ties to Chinese state-linked entities, U.S. procurement bans, and IP disputes. Analysts debate whether its “daily unlocks” strategy truly fosters open collaboration or disguises proprietary advantages, as geopolitical tensions intensify in the west-east AI rivalry.