What to know

  • OpenAI has launched Flex processing, a new API option that halves costs for o3 and o4-mini models.
  • Flex processing is designed for non-critical, low-priority, and asynchronous AI tasks.
  • Users trade lower prices for slower response times and occasional resource unavailability.
  • ID verification is now required for certain developers to access these models.

OpenAI has introduced Flex processing, a new API option that offers significantly reduced costs for using its o3 and o4-mini reasoning models. This option is aimed at developers and organizations who need to run lower-priority or non-production AI tasks, such as model evaluation, data enrichment, or asynchronous workloads. By choosing Flex processing, you accept slower response times and the possibility of occasional unavailability in exchange for a 50% reduction in API costs.

For the o3 model, Flex processing costs $5 per million input tokens and $20 per million output tokens. This is a substantial drop from the standard rates of $10 and $40, respectively. The o4-mini model sees its input token cost fall from $1.10 to $0.55 per million, and output token cost from $4.40 to $2.20 per million. These savings can make a significant difference for developers or businesses with limited budgets who do not require immediate results from their AI tasks.

Flex processing is currently in beta and is intended for tasks that do not demand real-time performance or guaranteed availability. OpenAI has positioned this offering as a response to the rising costs of advanced AI services and increased competition from other providers, such as Google, which recently launched its own budget-friendly AI models.

To access the o3 model under Flex processing, developers in usage tiers 1-3 must complete a new ID verification process. This requirement is part of OpenAI's efforts to prevent misuse and ensure responsible use of its technology. The company has stated that this verification step is necessary to maintain the security and integrity of its platform.

With Flex processing, OpenAI aims to make its AI models more accessible and affordable for a wider range of users, especially those working on experimental, background, or large-scale data processing tasks where speed is not the highest priority. As the AI landscape evolves, such options may become increasingly important for balancing cost, performance, and accessibility.

Via: TechCrunch