For a short (250 word) System Message, an assistant starts reponding in about 4 to 6 secods, while Chat almost responds immediately (gpt-4o-mini)
Is this expected?
I don’t use any specific assistant features, such as code interpreter or file search. To make sure it wasn’t the client, I tried same system and user prompts side by side on the playground, same results. Assistants very consistently wait a few seconds before start streaming.