DeepSeek unveils low-cost AI model intensifying global tech race

Chinese startup DeepSeek has introduced a new artificial intelligence model, DeepSeek-V4, promising significantly lower computing and memory costs while delivering advanced capabilities. The launch comes more than a year after the company disrupted the global AI landscape with its earlier low-cost reasoning model.

Based in Hangzhou, DeepSeek said the new model features an ultra-long context window of up to one million words, enabling it to process far larger volumes of information compared to most existing systems. The company described the model as “world-leading” in reasoning, knowledge, and agent-based performance.

The release highlights intensifying competition between China and the United States in artificial intelligence development. Recently, the White House accused Chinese entities of attempting to acquire US AI technology through large-scale efforts, further escalating tensions in the sector.

Pakistan’s critical minerals push gains momentum amid Reko Diq developments

DeepSeek-V4 is available in two versions: Pro and Flash. The Pro variant, with 1.6 trillion parameters, offers high-end performance, while the Flash version provides a more efficient and cost-effective alternative. Both versions aim to make advanced AI tools more accessible for commercial and research use.

Experts say the development could mark a turning point for the industry. By reducing the cost of long-context processing, the model could enable wider adoption of AI in sectors such as finance, healthcare, and enterprise solutions.

DeepSeek gained global attention last year when its chatbot, powered by the R1 model, matched leading systems like ChatGPT at a fraction of the cost. However, concerns around data privacy and content restrictions remain.

The company’s open-source approach continues to differentiate it from competitors such as OpenAI and Google, which rely on proprietary systems.