Temperature is the creativity dial.
Direct answer
Temperature controls how random a language model's output is. Low values make answers more predictable and focused, which suits factual and coding tasks. High values add variety, which suits brainstorming and creative writing. Many tools also offer top-p, which limits choices to the most likely words.
Rough guide
| Setting | Behaviour | Good for |
|---|---|---|
| Low | focused, consistent | extraction, code, facts |
| Medium | balanced | general chat |
| High | varied, surprising | ideas, stories |
How it works
The model calculates probabilities for the next word. Low temperature sharpens those probabilities so the likeliest word wins more often. High temperature flattens them so unlikely words get a chance.
Tips
- Change one setting at a time.
- High temperature does not make a model smarter, only more varied.
- For reliable output, pair low temperature with clear instructions.
Where Teyro fits
Teyro's The AI Stack course turns this into short lessons you can play in a few minutes a day. It has 100 lessons, and you can start free.
The bottom line
Use low for accuracy and higher for imagination.
