Home / Glossary / Temperature
What is Temperature?
Temperature controls the randomness of an LLM’s output — 0 is deterministic, higher values introduce variability. osFoundry exposes temperature per chat request and per agent profile.
Detail
Mathematically, temperature scales the logits before softmax — higher temperature flattens the probability distribution, making less-likely tokens more probable. At 0, the model always picks the top token. At 1.0, you get the model’s natural distribution. Above 1.5, output gets noisy.
For factual tasks (code, data extraction, structured output) use low temperature (0-0.3). For creative tasks (writing, ideation) use higher (0.7-1.0).
Related terms