## OpenAI Introduces “Flex Processing” for Cost-Conscious AI Developers
In a move to stay competitive against rivals like Google, OpenAI has announced the beta launch of “Flex processing,” a new API option offering significantly reduced pricing for AI model usage. The trade-off? Slower response times and the potential for occasional resource unavailability.
Flex processing is specifically designed for less time-sensitive tasks and is currently available for OpenAI’s recently released o3 and o4-mini reasoning models. OpenAI envisions developers utilizing Flex processing for “non-production” tasks like model evaluations, data enrichment, and asynchronous workloads where speed isn’t paramount.
The pricing reduction is substantial. Flex processing cuts API costs in half. For the o3 model, the price drops from $10/M input tokens and $40/M output tokens to $5/M and $20/M respectively. The o4-mini model sees an even more dramatic decrease, going from $1.10/M input tokens and $4.40/M output tokens to just $0.55/M and $2.20/M.
This initiative comes as the cost of cutting-edge AI continues to rise, prompting a surge in budget-friendly models from competitors. Google, for instance, recently unveiled Gemini 2.5 Flash, a reasoning model boasting performance comparable to DeepSeek’s R1, but at a lower cost per input token.
Beyond the price reduction, OpenAI is implementing stricter access controls. According to an email sent to customers, developers in tiers 1-3 (determined by spending on OpenAI services) will be required to complete a newly introduced ID verification process to gain access to the o3 model. This verification extends to reasoning summaries and streaming API support for o3, and potentially other models in the future.
OpenAI has stated that the ID verification process is aimed at preventing misuse of its platform and ensuring compliance with its usage policies. This move signals a greater emphasis on responsible AI development and deployment, alongside making its services more accessible to a wider range of users with varying budget constraints.