Due to the new and, frankly, user-unfriendly pricing policy, our Pro plan limits have been maxed out. As a result, I can no longer access premium models like Claude or Gemini.
After that, I browsed other threads on the forum and, following the suggestions made by the team, I tried using Auto Mode — which is supposed to work with premium models.
However, no matter how well we provide context to the model, it either fails to understand or simply refuses to carry out the task. Sometimes, it just gives a vague response and avoids actually completing what was asked.
Is this really what we’re supposed to call a “premium” model? From my point of view, there’s absolutely nothing usable about it in its current state.
Additionally, when using Auto Mode, we should be able to see exactly which model is being selected. Based on my analysis, it seems Cursor defaults to GPT-4.1 — a model that clearly doesn’t perform well in this context.
In short, I don’t think Auto Mode is usable at all right now.
I also tried to find clear information about exactly when our exhausted limits will reset so we can use the models again, but couldn’t find anything. I believe there should be clear and accessible information indicating exactly when our limits will refresh.
Here, they updated the docs with the new pricing details. Basically the limits are monthly, that’s it. You get 20$ of use at API cost, and once you reach the limit you get some more every 5 to 24hs.
Some people use Cursor for just a few quick tasks. But if I wanted something fast and simple, I would just use Claude directly. The reason we pay for Cursor is to handle more complex work. It’s not to have a low-quality model running in automatic mode.
This pricing policy is outright terrible. They need to either revert to the old system or come up with something actually fair and usable. At this point, the Pro versions are completely worthless. And the so-called automatic mode is absolute garbage.
