A probabilistic model makes decisions or predictions based on the likelihood of different outcomes rather than producing a single definitive answer. Most language models are probabilistic — given an input, they assign probabilities to possible outputs and generate from among the most likely ones.
This is why the same prompt can produce slightly different outputs each time, and it is why temperature settings matter for controlling how much variation the model introduces. For use cases that need high consistency, low temperature and structured output formats help narrow the range of what the model produces.
