Temperature controls the randomness of a model’s generated output. Google’s machine learning glossary describes higher temperatures as producing more random output and lower temperatures as producing less. The setting concerns variation in generation, rather than providing a measure of factual accuracy. Source: Google for Developers: Temperature

Test it against a clear task

Ask the supplier which settings the chosen model supports. Compare several outputs using the same approved prompt and record the setting used. For an enquiry summary, assess whether essential dates and requirements remain accurate, rather than judging only how fluent the text sounds.

Keep the factual review in place

Check names, prices and conditions against the supplied evidence at every setting. Agree who reviews the output before it reaches a traveller. This keeps the publication decision tied to verified content instead of treating a lower randomness setting as approval.

Written and updated by Travelgenix. Sources checked on 19 October 2026. Practical advice and examples are Travelgenix editorial guidance.