Chinese AI company DeepSeek has launched its latest open-source model, V4, marking a major step forward in large-scale AI development. The new model includes two versions: V4-pro, with 1.6 trillion parameters, and V4-flash, with 284 billion parameters. Both models support a 1 million token context window, significantly expanding how much information they can process at once while maintaining strong cost efficiency. DeepSeek claims the models are competitive with leading systems from US firms such as OpenAI and Google DeepMind. The release highlights China’s rapid progress in advanced AI capabilities, despite ongoing restrictions on semiconductor technology. At the same time, Huawei has pledged full support for DeepSeek’s development by providing new AI chips, reinforcing domestic efforts to reduce reliance on foreign hardware. This comes amid escalating US-China tensions over technology exports, with Washington advancing new measures to restrict China’s access to advanced chipmaking tools. The developments reflect a broader global race in AI and semiconductor technology, where model performance, compute infrastructure, and chip supply chains are increasingly central to national and corporate competition.
Read more here
