If I try to use models like Gemini 2.5 pro (which, by the way, does not support thinking on Kagi Assistant?), it often fails stating that it reached a timeout.
When comparing competitors of kagi assistant, on the same Gemini 2.5 pro (and even its thinking version which Kagi do not have) I can have a multiple message conversation with the model, while the first message is still being processed on Kagi Assistant.
Given that Gemini 2.5 Pro is currently one of the best model for programming, this is a bit frustrating 😅
