DeepSeek unveils new flagship AI model a year after breakthrough
DeepSeek unveiled a new flagship AI model, V4 Flash and V4 Pro, with upgrades and new features. The Chinese startup's previous model, R1, caused a significant stock market selloff in 2025.
DeepSeek rolled out preview versions of its new flagship artificial intelligence model on April 24, 2026. The V4 Flash and V4 Pro models were introduced with basic details on pricing, parameters, and output capacity. The Hangzhou-based AI startup previously released the R1 model, which rivaled top AI systems from companies like OpenAI at a lower cost. This caused a trillion-dollar stock market selloff in 2025 as some tech firms and investors reevaluated their AI investments. American technology giants are now projected to invest around $650 billion in 2026 on AI infrastructure and data centers. DeepSeek's new models have a maximum output of 384,000 tokens.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.