Anyone know when we can expect to see speed start to pick up for GPT-4?

Right now I’m seeing anywhere from 30 seconds to a minute on response times for the GPT-4 API with prompt tokens of ~1500 and response tokens of ~400. Does anyone know when we can expect to see these times start to speedup to sub 30 seconds? I have found little information on this on the web and other forums.

4 Likes