Hello guys!
I just saw that Altman confirms full o3 and o4-mini “in a couple of weeks”. Thats great, but both of these are reasoning models.
My question is, when do you think a predecessor to 4o-mini will come out?
Hello guys!
I just saw that Altman confirms full o3 and o4-mini “in a couple of weeks”. Thats great, but both of these are reasoning models.
My question is, when do you think a predecessor to 4o-mini will come out?
Yeah, would like to know this as well. The fact that this doesn’t seem like a bigger priority should be classified as a CRITICAL OUTAGE!
What’s missing is a gpt-4o-2024-05-13-pro that pays even more attention and doesn’t spit emojis.
Would you care to elaborate a bit more? Do you mean this is the best model?
As a chatbot dev that has clients with a lot of website traffic, the token cost is to high to only use 4o.
As a chatbot dev, many will avoid you if you only have the lowest-performing AI in your roster, without any ability to upgrade quality for a price.
There are three variants of gpt-4o. The initial release is twice the cost, and for developing custom applications, a switch to it is often all you need to get things to work properly that need high quality understanding and instruction following - and not to be merely ChatGPT. Give it the attention to an entire context that gpt-4.5 (orion) has, along with no inner curtailment of the lengths it will output, and it could definitely be a winner for utility.
Thanks for the advice.
In my experience, for most customer support or product recommendation chatbots, simpler interactions are the norm. And when you’re dealing with high-traffic websites, cost-efficiency becomes critical.
I only use 4o. I never use the reasoning models. What do you guys use them for? It seems like they are good at solving math questions but I don’t do math
In my opinion 4o mini is good enough for most tasks. At the end of the day 4o is 16.6 times more expensive. Is the quality so much better in 4o is the question, not sure.
my apologies, i’m referring to personal use. your probably talking about the API? If its 16 times more expensive, mini would probably be way to go, seems optimized to be efficient
No worries my friend! Yes correct im talking about the API. Thanks for you input.