In an attempt to strengthen its competitive position against rival AI companies like Google, OpenAI has rolled out Flex processing, a new API option that allows lower AI model usage costs in exchange for slower response times and possible intermittent unavailability of resources. According to OpenAI, currently in beta, Flex processing is aimed at OpenAI’s newly launched o3 and o4-mini reasoning models in “non-production” tasks and model evaluations.
Not long ago, OpenAI gave the idea of equipping Natural Conversation Technology, an advanced AI dialogue algorithm with the current AI model. With the new option, customers can get away with paying only half of the API fee for more complicated and potentially less urgent production-related tasks. The pricing for the Flex o3 version is $5 per million input tokens (approximately 750,000 words) and $20 per million output tokens, compared to regular pricing of $10 per million input tokens and $40 per million output tokens.
The price of Flex o4-mini is $0.55 per million input tokens and $2.20 per million output tokens, which is a lot less when compared to the pricing of $1.10 as well as $4.40 per million tokens. Flex processing was announced amid rising advanced AI technology costs, with competitors launching cheaper and better alternatives. Recently, Gemini 2.5 Flash, an upgraded reasoning model from Google, was launched; it either equals or outperforms DeepSeek’s R1 with a lesser charge per input token.
OpenAI has mentioned that the developers in the first three tiers will need to pass a new ID verification only then will they get access to the o3 model. It is a tier depending on the money they invest in the services of OpenAI.
The access for reasons-oriented summaries and streaming API support of o3 and other models will also need verification. OpenAI has previously stated that the ID verification intends to minimise misuse of the services against its interests and reputation.
