I have had this happen with me a few weeks back, albeit with a very different thing, their API for GPTv4-1106 (which I understand is a preview model but for my

I have had this happen with me a few weeks back, albeit with a very different thing, their API for GPTv4-1106 (which I understand is a preview model but for my use case,the higher context length was quite important which that model has). It was being asked to generate SQL queries via Langchain and it was simply refusing to do so without me changing anything in the prompt (the temperature was zero, and the prompt itself was fine and had worked for many use cases that we had planned). This lasted for a good few hours. The response it was generating was "As an OpenAI model, I cannot execute generate or execute queries blah blah) As a hotfix, we switched to the other version of GPT4 (the 0125 preview model) and that fixed the problem at the time.