Right now, for model selection, either I select “Auto” or one of the models. But most of us have 2-3 models we keep switching among to get the best out of Cursor and optimize the cost. I should be allowed to assign 2-3 models, e.g. “Claude 4 sonnet” and “Claude 4 Opus thinking” and then Curor should make the best judgement keeping in mind the complexity and cost trade-offs.
1 Like
i believe that is what the team is focusing on, they see a lot of people use expensive models, for obvious reasons, the cheaper ones just dont get the job done, so they are trying to steer us into the cheaper ones in order for cursor to be profitable, eventually…
however based on what i see they shall say that directly and give out clear limits of that people get out of subscriptions, not hiding the cost savings, e.g. the auto mode not telling you what model it uses
we all understand cursor cant just keep losing money forever by giving people unrealistic capacity, but i think we, the people who pay for it, should be well aware of what we do and what we dont get out of our subscriptions and adjust accordingly
first there was no limit, then there was a limit, then we got an extra usage on some models every now and then… i understand it is not easy to figure out what the company can afford to give out and might need to test that out in production to figure out some hard limits, but if that is the case, i personally would have preferred if the team would have let us known that they are conducting sort of a testing phase to see if this pricing model is viable and if it helps them manage their profitability moving forward, it is just that from our point of view, the behavior seems deceiving, even though, very unlikely the case..