Comment by TylerGlaiel

I think openAI might be experimenting with smaller context lengths to save on costs or something since I've had a few other things break down like this for me today too (even in GPT-4)