The only "limit" I have really with o3 is my patience -- it's a slow model for regular use. If I don't need its intelligence I'll use o4-mini or even 4o for speed, saving o3 for the prompts that really need it.
These days the only model I find I get rate-limited on frustratingly quickly is 4.5. It's clear OpenAI does not really want you using that one, despite the fact that's very good (and probably very expensive for them)! Pretty underrated imo.
For my money (my actual money!) o3 is still the best model I’ve used. That is included in the $20 a month plan.
I think you're overestimating what people use llms for. The only thing they're democratising is themselves
This is a weird conversation. At first you were concerned about prices and now you’re railing about the US Navy not using Chinese models. What’s your problem here?
Until you can run high quality models on affordable devices on your desk or in your hand the extent of the democratisation is much more limited than you might like.
Perhaps OSS will come to the rescue here.
(Aside: obviously free tiers are available but these are all hobbled in various ways: usage limits, data sharing/leakage, etc.)
Which part exactly ? The part where everyone pays 20+ a month to a few megacorps or the part where we willingly upload all our thoughts to a central server ?