Donât really use it, so I really donât knowâŚbut out of curiosity, does 4o-mini provide benefits over 4o? In regards to GlideAI, wouldnât Glide still be absorbing the cost, so if they are already offering 4o in place of 4o-mini in GlideAI, wouldnât that be a good thing?
Itâs less cost for Glide, but they also charge you updates based on how much you use it.
So if they eventually adopt GPT-4o mini, I guess it will be 1 update per [something between 1333 and 4096] tokens. I donât have the numbers immediately here to provide an accurate estimation.
In terms of intelligence, I donât think it would be as good as what we had before, specifically GPT-4o, but itâs faster & cheaper.
Got it. That makes sense.
The gpt-4o-mini would replace the standard model only (which probably uses the less capable and more expensive gpt-3.5-turbo). The High Power model and High Power v2 model would continue to be powered by gpt-4 and gpt-4o, respectively.
We use gpt-4o as the standard model for most Glide AI actions, so we are being careful about moving to mini.
We canât access the browser tool via Glide AI right? So no realtime internet search?
I canât speak for the development team of Glide unfortunately.
You would have a better chance building assistants and have a âweb searchâ function call.
Or just use something off the shelf like Perplexity.