Update agents to use optimizing completion API.

[mmirman]

This will require re-running of tests. It should be possible to turn off this behavior at a high level. Blocked by both #2 and #18

[cartazio]

Is this task done ?

[abhigya-sodani]

Not yet

[mmirman]

@abhigya-sodani I will note this isn't exactly straightforward as a lot of the agent calls use highly-varying prompts and the optimizing endpoints are meant to improve cases where the prompts don't vary between calls

[cartazio]

@abhigya-sodani is this done?

[abhigya-sodani]

I don't think we should do this. REBEL will not work with anything except GPT/Chat_GPT. Our completion API can be configured to be any LLM, and will cause REBEL to fail.

[cartazio]

so we should close this ticket?