OpenAI has released Flex processing, an application programming interface (API) option, aiming to outdo leading AI companies such as Google.
It offers reduced artificial intelligence (AI)-powered model usage prices in exchange for “occasional resource unavailability.”
Flex processing is currently available in beta for the ChatGPT manufacturer’s recently introduced o3 and o4-mini reasoning models.
It minimises API costs by half. For o3, Flex processing cuts down the price to $5 per million input tokens (~750,000 words) and $20 per million output tokens.
For o4-mini, Flex processing is $0.55 per million input tokens and $2.20 per million output tokens.
The release of Flex processing comes as the price of frontier AI continues to increase and competitors launch budget-oriented models.
It is pertinent to mention, OpenAI emphasises that developers can avoid timeout errors by raising the default timeout.
These APIs are automatically set to timeout at 10 minutes. However, with recently launched Flex processing, lengthy prompts may take longer time than that.
OpenAI stated that raising the timeout will minimise the chances of errors.
In an email to customers revealing the release of Flex pricing, OpenAI suggested that developers in tiers 1-3 of its usage tier hierarchy will be required to complete the latest ID verification process to access 03.
Previously, OpenAI stated that ID verification aims to prevent bad actors from violating its usage policies.
