Chinese artificial intelligence startup DeepSeek on Friday, April 24, 2026, announced the release of its long anticipated new AI model, more than a year after it drew global attention with a low cost reasoning system that rivalled leading United States technologies.
The company disclosed that its latest model, DeepSeek V4, features an ultra long context capacity of up to one million words, positioning it as a highly competitive and cost effective solution in the rapidly evolving AI space.
According to the company, the new model delivers strong performance across key areas, including reasoning ability, world knowledge and agent based capabilities, placing it among the leading open source AI systems globally.
A preview version of the model has been made available, with two variants released DeepSeek V4 Pro and DeepSeek V4 Flash. While the Pro version is designed for more advanced tasks, the Flash version is described as a more efficient and economical option due to its smaller parameter size.
DeepSeek said the V4 Pro model operates with 1.6 trillion parameters, while the V4 Flash version has 284 billion parameters, both aimed at improving decision making and task execution. The model has also been optimised for use with several AI agent platforms, enhancing its application across coding and automation tools.
The company added that in benchmark tests, the V4 Pro model outperforms most open source systems and comes close to matching top tier closed source models such as Google’s Gemini Pro.
DeepSeek, based in Hangzhou, first gained global prominence in January 2025 with the launch of its R1 reasoning model, which powered a chatbot that performed at levels comparable to platforms such as ChatGPT but at significantly lower development cost.
The breakthrough, widely referred to as the “DeepSeek shock,” triggered a sell off in AI related stocks and prompted a reassessment of industry strategies, challenging assumptions about United States dominance in artificial intelligence.
Despite its rapid rise, the company has faced scrutiny over data privacy and content restrictions, with its chatbot reportedly declining to address certain politically sensitive topics.
In China, DeepSeek’s tools have seen growing adoption across sectors, including healthcare, finance and public administration, partly driven by its open source approach, which allows developers to access and build on its systems.
The latest release comes amid intensifying competition between China and the United States in artificial intelligence development, with both sides increasing investments and raising concerns over technology security and global influence.
Source: AFP
