Assistants getting "cold"?

Hey community,

Lately, I’ve got a feeling that an Assistant gets “cold” after some time of inactivity.

To put it clearer, during mornings, when users start interacting with an Assistant, initial responses require long times to be generated and, after a while, these times become shorter.

Has anyone else experienced this as well?

I’m not sure if it’s related to the fine-tuned model that relies within an Assistant, or just the API itself (maybe due to giving priority to other organizations that are generating more traffic ¿?)


1 Like

It’s possible that your observation is related to how the Assistant’s underlying model behaves after periods of inactivity. Some models may take longer to generate responses initially but become more responsive over time as they “warm up” from increased usage. It could also be influenced by factors such as server load or resource allocation within the API.

1 Like