In my view, the three new models are extremely disappointing. Changes and fixes are implemented only sporadically, and while VMax is being presented as being just as good as Opus—or even “Opus”—Vercel states that it is not the same model as 4.5. On top of that, credits are being consumed at an alarming rate, and many requested changes simply aren’t being made anymore.
Because of this, I’ve reluctantly decided that I will only continue using Vercel once Opus 4.5 is available again. This is very frustrating, and I’m surprised that more people haven’t raised concerns about it yet.
Have you tried just using the Pro model and see how that one works for you?
I havent personally experienced any issue with the Max (Opus model). I thought it worked fine as it did before. Is there anything else maybe going on like file (chat) size or something?
Very curious to be honest what may be causing this experience.
LLMs do not have innate knowledge about what their model label is, they will just guess based on training data unless the system prompt explicitly tells it to act as a certain model, which the v0 agent’s prompt does not
