nicoty If they were to do so, it would probably be extremely rate limited. At the moment, the RPM (requests/minute) is at 20 for o1-preview
. Keep in mind, Kagi is probably using just 1 (or a couple) OpenAI keys to fulfill requests to the assistant, so this 20 requests/minute would be affecting everyone across Kagi.
I would like to see Kagi implement this feature where you get a few requests per day for now, given the API limits, however, since it just looks like a simple drop in change for the model on API code.
Details: https://platform.openai.com/docs/guides/rate-limits/usage-tiers?context=tier-five