BEIJING, China — DeepSeek released the DeepSeek-V4 artificial intelligence model on Friday, April 24, 2026, featuring an ultra-long context of one million words.
Okay News reports that the company announced the system is cost-effective in a statement on the social media platform X. The model is available in two versions, with DeepSeek-V4-Pro utilizing 1.6 trillion parameters and DeepSeek-V4-Flash using 284 billion parameters.
The company said the model is optimized for products including Claude Code, OpenClaw, OpenCode, and CodeBuddy. A statement from DeepSeek claimed the model achieves leadership in domestic and open-source fields regarding agent capabilities, world knowledge, and reasoning performance.
The statement added, “In world knowledge benchmarks, DeepSeek-V4-Pro significantly leads other open-source models and is only slightly outperformed by the top-tier closed-source model, (Google’s) Gemini-Pro-3.1.”
The release coincided with reports that Meta plans to cut a tenth of its staff while increasing investment in artificial intelligence, and reports that Microsoft is looking to trim its ranks. The White House on Thursday, April 23, 2026, accused Chinese entities of attempting to steal artificial intelligence technology. Michael Kratsios, the science and technology chief for the United States, said, “The US has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI. We will be taking action to protect American innovation.”
Chinese Premier Li Qiang stated at a gathering of top decision-makers last month, “China-made large AI models spearheaded the development of the global open-source AI ecosystem.”
