DeepSeek announced via its official WeChat account this evening (the 13th) that the official version of DeepSeek-V4-Pro is now available across the app, web, and API. Users can select “Expert Mode” in the app or on the web to use the new official V4 Pro model.
The official version of DeepSeek V4 Pro features enhanced agent capabilities, with particularly significant performance improvements in production environments.
The DeepSeek API now natively supports the OpenAI Responses API format and has been specifically adapted for Codex. According to the announcement, users can refer to the official API documentation: Getting Started – Integrating Agent Tools – Codex, and configure Codex using the one-click setup script.
The V4-Pro and V4-Flash thinking modes now support three levels of thinking intensity: low, high, and max. Users can select an option based on task complexity: low for simple tasks, high for everyday agent tasks, and max for highly complex scenarios. For setup instructions, refer to the official API documentation: API Guide – Thinking Mode.
Following the official release of the entire DeepSeek V4 model lineup, DeepSeek will update its API pricing using a “peak and off-peak pricing” mechanism: off-peak prices will be half of peak-hour prices, encouraging users to adjust task timing based on actual usage. The new prices will take effect at 0:00 Beijing Time on August 17, 2026.
