Very interesting, thanks for these tests Jim! I've been doing my own tests and arriving at similar conclusions as you. There is another 'level' of testing I've been experimenting as well and will publish my anecdotal results very soon. I have a feeling that despite best efforts and all the money in the world, it'll be hard for anyone to catch-up with OpenAI.... they've been building and testing their GPT models for YEARS and they will continue to lead for some time. Interesting times we live in, no?