[_boffin_]

It seems like it's running at comparable speeds to GPT4 prior to Turbo. I could be wrong, but what I'm trying to say, it ain't bad at all.