Aug 21, 2025 05:08 PM (GMT+8) · EqualOcean
According to the official WeChat public account of DeepSeek, DeepSeek-V3.1 has been officially released. This upgrade includes the following main changes: it features a hybrid reasoning architecture, which allows one model to support both thinking mode and non-thinking mode at the same time; it has higher thinking efficiency—compared with DeepSeek-R1-0528, DeepSeek-V3.1-Think can provide answers in a shorter time; it has stronger Agent capability—through Post-Training optimization, the new model has achieved significant improvements in performance when it comes to tool usage and agent tasks.

Source: