The new Cursor 2.0 model selection no longer shows whether it is a thinking model or not. Previously you knew whether you were on a thinking or non-thinking model just at a glance. In 2.0, the brain icon is not shown next to the model, so I am constantly having to click and see if I’m using a thinking or non-thinking model (depending on the chat, I use one or the other as I feel most appropriate, so its never the same model, never even the same type of model…it could be Claude or something else.)
Steps to Reproduce
Select thinking model.
Expected Behavior
Selected model display should always show whether it is thinking or not by showing the brain icon for thinking models.
Operating System
MacOS
Current Cursor Version (Menu → About Cursor → Copy)
i dont think that it is a bag. I think they took it out because there was not enough space because of additional icons like language browser audio
so you just nned be sure which model you use and that’s all
Right, and because I use multiple models, I’m constantly having to check and make sure that I’m usnig the right model now, because a thinking model is very different from a non-thinking one in many ways.
Seems to be plenty of space. At the very least, if there is plenty of space, they should show it, and only hide the brain if they are literally running up against the icons on the right-hand side.
The chat window no longer shows an indicator for whether the currently selected model is a thinking or non-thinking variant.
In other words, there is no visual distinction between thinking and non-thinking variants of the same model. Sonnet 4.5 and Sonnet 4.5 thinking are both displayed as just Sonnet 4.5 in the chat box when “currently selected”.
Steps to Reproduce
Open a new chat.
Select a thinking model from the dropdown (especially one like Sonnet 4.5, which has both a thinking and non-thinking variant).
Look at the model name in the chat box. It won’t have the little brain icon, or any other “thinking model” indicator.
Expected Behavior
I expect some visual distinction to be present for thinking models. Older versions of cursor did this by putting the little brain icon next to the model name. Doing the same for 2.0.x and above would be ideal.
I switch between thinking and non-thinking Claude models very frequently. Having no visual distinction between the two model types is making me second-guess myself multiple times in the middle of each chat, which is frustrating, especially since this wasn’t an issue on earlier cursor versions. A fix would be much appreciated!
I just noticed MAX mode has the same issue as well. In 1.7.x and earlier, the “currently selected” model in the chatbox would become the purple/blue gradient to show when MAX mode was selected. However, that is also no longer the case in 2.0.x.
once you select an agent, it doesnt show in the input box if its the thinking or non-thinking version, which means you have to hit the pulldown to check. And if you misclick you wont know. Terrible UX. Please revert this. (There’s a reason you show it in the menu list, the same reasons apply here…!)
Steps to Reproduce
Select a thinking agent and realize you dont know if you selected the right one.
Expected Behavior
The previous versions of cursor for 12-18 months.
Operating System
MacOS
Current Cursor Version (Menu → About Cursor → Copy)
@deanrie Any word on this? We are now up to Cursor 2.2…this was broken with Cursor 2.0. I can’t imagine it is a difficult feature to fix, and I’m baffled at why this is not being resolved. There still seems to be this endless march towards new features…and every time new features are released, they are buggy, often intorduce non-trivial problems, and they are left to rot.
A GOOD product, has a BALANCE of new feature work AND refinment, debugging and optimization of EXISTING features. Cursor needs to put some effort into stabilizing and debugging their existing feature base. So many features have been fundamentally broken or barely usable for a long time now, many months, while new half-baked features keep getting shunted in release after release (only to ALSO be forgotten by the next release…)
Please, support your EXISTING customer base, and fix the plethora of broken or incomplete or unstable features soon.
This is one of many major feature-REMOVALS you guys did with the original 2.0 release. Your users have seen little to no progress on RESTORING the losst functionality. This includes other things as well, such as the loss of the ability to pop the agent panel out into its own window so it can be moved to another screen.
It really seems like you guys created a branch for 2.0 work, worked on it as a major tangent, then NEVER RE-INTEGRATED the changes being done in versions 1.5, 1.6, 1.7, 1.8, BACK INTO the branch for 2.0… So many of the innovations that were RECENTLY INTRODUCED in 1.7 and 1.8 for example, like the ability to pop out the agent panel into its own window, were just…gone. We need this functionality…it saves us time, allows us to be more efficient by effectively using our screen space (i.e. popped out agent panel), etc.
