Prompting
Temperature
A setting that controls how random or creative an AI model's responses are — lower values produce focused answers, higher values produce more varied ones.
01 ——
In plain English
Temperature is a numerical parameter (usually 0 to 1, sometimes up to 2) that controls how much randomness the AI uses when picking the next token. It's the main creativity vs. consistency knob.
What different temperatures do:
- 0.0 — deterministic, picks the highest-probability token every time. Same input → same output. Best for code, classification, structured data.
- 0.3–0.7 — moderate variation. Good for chat and most tasks.
- 1.0+ — high randomness. Better for brainstorming, creative writing, generating diverse options.
Common pitfalls:
- Setting too high → incoherent, off-topic responses
- Setting too low → repetitive, robotic outputs
- Comparing models at different temperatures → invalid benchmark
Most APIs default to around 0.7. For production tools that need reliability, lower it; for creative work, raise it.
02 ——
Related terms
Prompt Engineering
The practice of crafting inputs to an AI model carefully to get better, more reliable outputs.
LLM
Large Language Model — the type of AI behind tools like ChatGPT and Claude, trained to understand and generate text.
System Prompt
The high-level instructions given to an AI model at the start of a conversation that define its role, behaviour, and constraints.
Inference
The process of running a trained AI model to generate a response — as opposed to training the model.
Back to glossaryLast reviewed May 2026