Support OpenAI Flex Processing (and Maybe Batch Too).
OpenAI released o3 and o4-mini with the option to reduce cost by specifying flex processing. https://platform.openai.com/docs/guides/flex-processing
Please add an option to use flex processing.
OpenAI released o3 and o4-mini with the option to reduce cost by specifying flex processing. https://platform.openai.com/docs/guides/flex-processing
Please add an option to use flex processing.
Log in to comment and vote
Comments4
Amaranth Dew
Jul 1
Did anyone find a solution for this? I’m still interested.
Amaranth Dew
Oct 7, 2025
I would also like the option to use the batch API or flex tier for models that permit it (e.g., similar implementation for OpenAI, Anthropic). Even for smaller requests that are non-urgent, this would be a great option for additional cost savings. I agree with Aqua Toast that the best implementation would be a toggle in the chat.
Rose Pudding
Apr 23, 2025
batch can be hard, as it can be up to 24h and you need to harvest it instead of normal chat completion api
On the other hand, flex tier just need a simple body field
{"service_tier": "flex"}, still use the same api, just I little bit more wait time for 50% the price, it's a great deal. I really hope this can be supported to save money.Aqua Toast
Apr 19, 2025
Commenting for support. The 50% discount is meaningful.
I would appreciate an option to set a default 'service_tier' parameter or flex mode, but have a toggle option within the chat interface to switch from the default (like thinking for Claude 3.7).