Chinese artificial intelligence startup DeepSeek has unveiled its latest AI system, DeepSeek-V4, introducing a powerful model designed to significantly reduce computing and memory costs. The announcement arrives as competition intensifies between China and the United States in the rapidly evolving artificial intelligence sector, where technological leadership increasingly carries strategic and economic importance.
The Hangzhou-based company first gained global attention after releasing its R1 reasoning model last year, which powered a generative chatbot that rivalled leading American systems. With the release of DeepSeek-V4, the firm claims it has further advanced efficiency while maintaining high-level reasoning capabilities that can compete with major Western AI platforms.
Read more : Chinese AI startup Manus gains momentum with government support
According to the company, the new model features an ultra-long context capacity of one million tokens, allowing it to process significantly larger volumes of information in a single task. This capability places it alongside advanced systems such as Gemini developed by Google, marking a major step forward in long-text processing and complex AI operations.
The system has been released in two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, offering different levels of performance and efficiency. V4-Pro contains approximately 1.6 trillion parameters designed for advanced reasoning tasks, while the Flash version uses 284 billion parameters to deliver a more economical and energy-efficient alternative for broader commercial applications.
Read more : Trump’s AI ambition, China’s DeepSeek overshadow AI summit …
Industry analysts say the launch represents a significant turning point for artificial intelligence development because it addresses long-standing challenges linked to computing power and operational costs. Experts believe that ultra-long context support could soon move beyond research laboratories and expand into mainstream business sectors such as healthcare, finance, and public services.
Meanwhile, the release arrives during heightened geopolitical tensions between the United States and China over technological dominance. Officials in Washington recently accused Chinese entities of attempting large-scale AI technology theft, an allegation that Beijing strongly rejected, describing it as a baseless claim targeting the achievements of China’s growing artificial intelligence industry.