DeepSeek intensifies AI competition with V4 preview release. The Chinese startup launched its latest large language model, offering strong performance at lower inference costs than previous versions. The move comes over a year after DeepSeek's R1 model shocked global markets by achieving comparable results to leading AI systems while costing less than six million dollars to develop. V4 is available in pro and flash versions, optimized for agent-based tasks and knowledge processing. While analysts expect less market disruption than R1 generated, the release signals continued Chinese dominance in cost-efficient AI development. DeepSeek's open-source approach allows developers worldwide to access and modify the technology, intensifying pressure on global tech companies' massive AI infrastructure spending strategies and challenging the traditional U.S.
