Depends.

Your session is unique to you, and apart from intentional training by OpenAI the models do not learn. The session context is limited to 8000 tokens (roughly 1-2 per word) and that's the extent of its short-term memory.

I don't know what they have the temperature of ChatGPT set at, but there's a certain amount of random wiggle room within the probability tree if they have it set anywhere north of 0.