After an extensive development phase, DeepSeek's most formidable model has finally been unveiled. On the evening of August 12th, DeepSeek's official website updated its API documentation, elevating the model version from the V4-Pro preview to the official release, named "DeepSeek-V4-Pro-0813". This iteration boasts a context length capacity of up to 1 million Tokens, with a maximum output of 384K Tokens. Moreover, it seamlessly integrates with a wide array of mainstream Agent development tools and features. Regarding pricing, DeepSeek maintains its commitment to affordability; however, the official website has cautioned that API service pricing will undergo a substantial hike in the foreseeable future.