OpenAI to compete more seriously with companies like Google, a new option called “Flex processing»(Flex Processing) for API It has iroduced its artificial ielligence models. This option, currely provided in the beta version, allows the use of lower cost models, although it will be lower, and sometimes resources may not be temporarily available.
Flex processing is currely available for the new OpenAi reasoning, O3 and O4-Mini. This option has also been suggested more for low -cost or non -productive work such as modeling, data enrichme and time -consuming processing.
Reduce the cost of using OpenAI models
Cost of using API with this option exactly to Half Reduce. For the O3 model, Flex Processing Price is $ 5 per million input token and $ 20 per million token; While the normal price is $ 10 and $ 40, respectively. For the O4-Mini model, the price will be reduced to $ 1.50 per million input tokens and $ 4.40 for outbound tokens, respectively, $ 0.55 and $ 2.20, respectively.
This new option has been iroduced by OpenAI, which is increasing the cost of using advanced models of artificial ielligence, and competitors like Google also iroduce less expensive and optimized models. For example, Google has recely unveiled the Gemini 2.5 Flash model, which is less expensive or better than the Deepseek R1.
Openai sends an email to its customers to iroduce the Flex Processing option that users should complete the company’s new autheication process to access the O3 model (based on the cost of consuming) to access the O3 model. Support for the Summary of the Stream and API is also subject to this autheication.
Earlier, the company had stated that the aim of the autheication was to preve abuse and violate the policies of the company’s services.
