DeepSeek launches the V4 Pro, a 1.6 trillion-parameter model, as it accelerates efforts to compete with US rivals through technical improvements and strategic business moves amidst evolving API pricing and funding initiatives.
DeepSeek has released a new version of its flagship model, V4 Pro, in a move that sharpens the contest between Chinese and US artificial intelligence groups. According to the company, the V4-Pro-0813 release is designed to improve agent-style functions and is available through its application programming interface, mobile app and web platform.
The model sits alongside a broader V4 family that also includes V4 Flash. Reporting from Reuters said DeepSeek has altered the API pricing for both products, with separate peak and off-peak rates now in place. That shift suggests the company is trying to balance heavy demand with the high cost of serving large-scale AI workloads.
DeepSeek’s latest launch builds on the attention it drew earlier this year with its R1 model, which reignited debate over whether advanced AI systems can be built at far lower cost than many Western rivals. TechCrunch reported that V4 Pro is a 1.6 trillion-parameter model and that both V4 variants support a one-million-token context window, allowing them to handle very long documents and large codebases.
Independent assessments have painted a more mixed picture of the company’s progress. In May, the US National Institute of Standards and Technology said its CAISI evaluation found DeepSeek V4 to be the most capable Chinese AI model tested, but still roughly eight months behind leading US models overall. The institute said the system performed strongly on some benchmarks, including cost efficiency, even as the gap remained visible in areas such as reasoning, software engineering and cyber tasks.
DeepSeek is also trying to turn technical visibility into a more durable business. Reuters reported in July that the company was preparing a new funding round at a valuation of about $74bn, after previously raising around $7.4bn in its first outside round. It is also said to be expanding hiring for AI agent, data centre and chip design roles, as it seeks to reduce reliance on suppliers such as Nvidia and Huawei while competing more directly with OpenAI and Anthropic.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
