Hello,
I’m a user primarily from the API Playground for python coding , yet I face same issues with ChatGPT user which the model refused to write in full code block.
I mean, I’m willing to pay as per token count, it’s so ridiculous to have half-baked answer from the model.
It will always reply with 1 example pattern then expect me to follow the pattern which sometimes like 80% more works to write manually??
I pay to save my time not for actually manually writing for myself.
I know this maybe justified for ChatGPT user as load balancing inference, but for API user which pay by token generated?