DeepSeek launches V4 Pro at prices up to 14 times higher than V4 Flash

# DeepSeek Launches Costlier V4 Pro as AI Race Shifts Toward Advanced Agents

Chinese artificial intelligence company DeepSeek has launched its V4 Pro model, introducing significantly higher pricing as it targets more advanced AI applications. The new DeepSeek-V4-Pro-0813 model is priced at $1.32 per million input tokens and $3.96 per million output tokens, according to Artificial Analysis. The new rates are roughly nine times higher for input and 14 times higher for output compared with DeepSeek’s V4 Flash model. DeepSeek said the latest model delivers substantial improvements in AI agent capabilities, allowing systems to use tools, execute code and handle complex multi-step tasks with less human intervention. The model is now available through DeepSeek’s API, application and web products. The company also announced changes to its API pricing structure, including separate peak and off-peak rates for V4 Pro and V4 Flash. DeepSeek’s move represents a notable shift from its earlier strategy of emphasizing extremely low-cost AI access. Its V4 Flash model had attracted significant attention for offering advanced AI capabilities at a fraction of the cost charged by several international competitors. The launch of V4 Pro, however, places a higher premium on performance and agent-focused capabilities. DeepSeek said its V4 Pro API has also been updated to support the OpenAI Responses API format directly, while adding built-in integration with Codex. These features are designed to make it easier for developers to incorporate the model into existing AI-powered applications and coding workflows. Despite the price increase, DeepSeek’s new rates remain competitive with several major AI providers. The company had previously offered a 75% promotional discount on V4 Pro, making the latest pricing changes a significant reversal from its earlier low-cost approach. DeepSeek emerged as a major force in the global AI race in 2025 after gaining attention for developing powerful models with comparatively low computing costs. Its progress challenged assumptions about the amount of capital and hardware required to build competitive AI systems. Since then, competition among Chinese AI developers has intensified, with companies such as Alibaba and Moonshot expanding their own model offerings. The latest launch comes as DeepSeek also faces increasing financial demands associated with developing increasingly powerful AI systems. Reuters previously reported that the company was considering a new fundraising round at a valuation of around $74 billion. That followed its first major outside financing round, which reportedly raised about $7.4 billion in June. The potential fundraising effort highlights the enormous capital requirements of remaining competitive in the AI industry. Advanced models increasingly require access to high-performance computing chips, large-scale data centers and specialized engineering talent. DeepSeek’s decision to charge more for its most capable model could therefore reflect a broader shift in the economics of AI development. As companies move beyond basic chatbots toward autonomous agents and sophisticated coding systems, performance, reliability and integration are becoming increasingly important alongside price. The V4 Pro launch signals that DeepSeek is seeking to compete not only on affordability but also on advanced capabilities as the global AI race enters a more demanding phase.