TL;DR: I would never recommend this new AI for anything other than coding.I don't like being limited to 50 messages per week on o1, so I always used o1-mini (50 messages per day for Plus users), unless it couldn't do what I want.
R1 is on the same level as o1 (a little worse, however, I like its human-like chain of thoughts), but it has the same message limit per day as o1-mini, so now I use R1 instead of o1-mini, and o1 when R1 fails (but there are only a few things that R1 can't do, and o1 can). Since the input size is the same (128k tokens), I don't need to take this into account.
From what I understand, in a few days,
Plus users will have more messages for the o3-mini than the o1-mini allows now, and the o3-mini is supposed to have the same capacity as the o1 and R1, but it's faster.
Because of this, I think I'll stop using R1 soon.Edit: Actually, apparently there will be 100 messages per week (which corresponds to around 14.2 messages per day).
About the V3... I have no reason to use it. I only use advanced models like R1 and o1 when I want or need to program something (I don't know how to program yet, so I use AI for Python, CSS, HTML, and JS); if I want to do anything else, ChatGPT is better as it has a lot more options (I mean, look at this:
https://www.youtube.com/watch?v=iXXvdffzxzA). Ok, this example in the video was bad, since I usually don't use the advanced voice model much. What I meant is that this new AI has almost nothing but 4 buttons: use R1, search the web (that in 70% of cases it's better to use perplexity), send message, and edit message.
90% of my AI uses are not programming, but when it is, I spend all day on some project, which takes up a lot of messages, and
it's in this specific case that I use R1.