DeepSeek releases official V4 Pro model, raises API pricing for V4 Pro and V4 Flash, introduces peak and off-peak pricing

DeepSeek has officially launched its V4 Pro model, significantly enhancing agent capabilities and expanding its API offerings. The V4-Pro-0813 is now available through API, APP, and Web channels, marking a strategic move to regain market share against competitors like Moonshot AI and Alibaba.¹²

In addition to the product launch, DeepSeek announced plans to raise API pricing for both the V4 Pro and V4 Flash models, introducing peak and off-peak pricing to optimize revenue. This pricing strategy reflects the company's response to increasing competition in the AI sector, where it has faced challenges from rivals that have rapidly advanced their technologies.

DeepSeek's rise to prominence began with its R1 model, which went viral in early 2025, prompting discussions about the feasibility of developing powerful AI systems at lower costs compared to U.S. counterparts. The company is now valued at approximately $74 billion following a recent fundraising round that raised $7.4 billion in June. To support its growth, DeepSeek plans to double its staffing across various departments, including data-center and AI-agent teams, and is actively hiring chip-design engineers to develop proprietary AI chips, reducing reliance on suppliers like Nvidia and Huawei.³⁴