In the evening of August 13, DeepSeek made an official announcement regarding the launch of DeepSeek V4 Pro, along with concurrent updates to its mobile application, web-based platform, and API. The upgraded version boasts enhanced Agent functionalities, compatibility with OpenAI Responses API and Codex integration, and provides users with three distinct thinking modes: low, high, and max. Concurrently, DeepSeek revealed a modification to its API pricing strategy, introducing a peak-valley pricing system. Under this new system, prices during off-peak periods will be reduced to half of those charged during peak hours. The revised pricing scheme is set to come into effect at 0:00 on August 17. Specifically, during peak times, the cost for V4 Pro input (with cache hit) is set at 0.3 yuan per million Tokens, input (without cache hit) at 9 yuan per million Tokens, and output at 27 yuan per million Tokens.