We've been struggling with costs because our application chains together multiple calls to GPT to generate the output we want, and it was starting to be ~$0.08 per call which obviously isn't feasible for high volume applications.

This just made our business way more viable overnight lmao