On January 10th, the Information reported thatDeepSeek A new generation of flagship AI model, DeepSeek V4, will be launched in mid-February this year, during the new calendar year, and will be better able to write codes。
Internal tests indicate that its AI programming performance is expected to outperform industry leaders, including OpenAI GPT and Anthropic Claude。
According to the source, DeepSeek V4 has achieved a breakthrough in processing ultra-long code tips, which is a potential advantage for developers working on complex software projects。
The DeepSeek V3.2 model, which was released last December, is better than OpenAI GPT-5 and Google Gemini 3.0 Pro in some of the benchmark tests, and the V4 model exposed is expected to achieve further breakthroughs. 1AI follow-up will remain focused。