Feature Request: Deterministic Answer Option for Unit Testing

Getting some kind of switch to force a deterministic answer at the OpenAI API would be my preferred solution.
As a workaround, I will meanwhile delegate the answer comparison to the OpenAI API, in the sense of

Compare two texts in terms of content. If they are the same in content, answer "True", otherwise "False".

Text 1: "..."

Text 2: "..."
1 Like