The company’s V4-Pro-0813 model is priced at $1.32 per million input tokens and $3.96 per million output tokens, according to independent AI benchmarking firm Artificial Analysis.
The pricing represents a substantial premium over DeepSeek’s V4 Flash model, with the V4 Pro input price roughly nine times higher and its output price about 14 times higher. Artificial Analysis lists V4 Flash at $0.14 per million input tokens and $0.28 per million output tokens.
DeepSeek said the new model delivers significant improvements in capabilities designed for AI agents and is available through its API, mobile application and web products.
The company also announced changes to its API pricing for both V4 Pro and V4 Flash, including the introduction of separate peak and off-peak rates.
V4 Pro shows stronger benchmark performance
The higher price comes after an unusual development in DeepSeek’s recent model rollout.
Its V4 Flash model, released last month, unexpectedly outperformed an April preview version of the more advanced V4 Pro in several independent tests. The result raised questions about whether the company’s supposedly flagship model could deliver a meaningful improvement over its lower-cost alternative.
The latest independent assessments, however, suggest that DeepSeek has made substantial progress between the preview and official V4 Pro release.
Artificial Analysis awarded the reasoning version of V4 Pro a score of 53 on its Intelligence Index, compared with 40 for V4 Flash.
The index measures performance across nine areas, including agentic tasks, tool use, coding, scientific reasoning and the ability to work with long-context information.
The improved scores could help DeepSeek justify the significantly higher price of its Pro model, particularly among businesses and developers seeking stronger performance for complex AI applications.
DeepSeek seeks to regain its early advantage
The launch comes at a crucial point for DeepSeek, which became one of China's most prominent AI companies after its R1 reasoning model went viral in early 2025.
The model sparked global debate over the cost of developing advanced AI systems and challenged assumptions that cutting-edge models necessarily require the enormous levels of spending associated with leading US technology companies.
However, DeepSeek's early momentum has faced increasing pressure from a growing group of Chinese competitors, including Moonshot AI, Zhipu AI, MiniMax, Alibaba and ByteDance.
The growing competition has made it increasingly important for DeepSeek to demonstrate that it can maintain a technological edge while building a sustainable commercial business around its models.
Rising costs push DeepSeek towards external funding
DeepSeek is also facing the financial realities of competing in the AI industry, where developing increasingly sophisticated models requires substantial investment in computing infrastructure, advanced chips, data centres and specialised employees.
Reuters reported in July that DeepSeek was planning a new fundraising round that could value the company at around $74 billion. This followed its first external financing round in June, in which the company reportedly raised about $7.4 billion.
The move represented a significant shift for DeepSeek, which had historically operated without relying heavily on external capital.
The potential fundraising also highlights the enormous resources now required to remain competitive as AI models become more capable and rivals accelerate their development programmes.
DeepSeek has said it plans to at least double its workforce across several departments, including teams focused on data centres and AI agents.
The company has also been expanding efforts to develop its own artificial intelligence chips. Reuters reported in July that DeepSeek had stepped up private recruitment of chip-design engineers as it explores ways to reduce its dependence on suppliers such as Nvidia and Huawei.
With V4 Pro, DeepSeek is therefore attempting to address two challenges at once: maintaining its position in the rapidly evolving AI race and developing products capable of commanding higher prices.
If the model's stronger benchmark performance translates into greater demand from developers and businesses, the premium pricing could mark an important step in DeepSeek's transition from an AI industry disruptor into a more commercially established technology company.