Pretty sure that's exactly what LLMs in coding harnesses are.
Even when they implement LLMs.
I know ChatGPT seems like it's non-deterministic - but that just a user preference.
The only real issue is if you have many people on the same server, the GPU contention can be non-deterministic in rare cases.
LLMs can be deterministic if you need them to be - it's just that most people prefer the human-like interface.
Pretty sure that's exactly what LLMs in coding harnesses are.