OpenAI has introduced flex processing mode as a middle ground option between regular API calls and batch processing, which could take up to 24 hours. Additionally, it utilizes APIs like Chat Completions API and Responses API as usual.
When using flex processing, users can specify additional waiting time or default to 10 minutes. The connection will wait for processing to complete, making code adjustments minimal compared to regular API calls. However, if OpenAI doesn’t have available servers within the specified time, a 429 Resource Unavailable error will be returned.
The key advantage of flex processing is its cost-effectiveness, similar to batch processing, allowing tasks that don’t require immediate responses to be more efficient. Currently, it is only available for models o3 and o4-mini.
TLDR:
OpenAI introduces flex processing mode as a cost-effective option between regular API calls and batch processing, providing additional waiting time for tasks to complete efficiently.
Leave a Comment