I almost dont even need to say anything this should be obvious how ridiculous this is.
i don’t want glm 4.5 to be a cursor/auto model
this is not what my wording said, this is not what I intended, I actually said
”GLM 4.5 for default auto model”, meaning that you dont randomly change which model auto defaults too, you dont just make it another cursor model, you make auto mode actually by default have a huge massive upgrade.
you just blanket on every post “if this gets enough upvotes we will consider it” this is ****, you guys wont consider making cursor better, just to make the product better even when it is totally within your scope to do, its not even difficult to do.
**
LOL Cursor that’s it for me I am out this month, with or without refund (u choose wisely lol)
You can see it is compressing/truncating my context files (EVEN THE DATABASE schema.ts FILE !!!) even in MAX MODE… and they remove the Manual Edit mode so you see their “Ask” mode wasted a series of 8 x 250k requests (that is $1.5 ) where normally 3 weeks ago before their sneaky silent changes I would first use Gemini 2.5 Pro in MAX MODE in Manual Edit with my full codebase of 300k tokens, then use Gemini 2.5 Pro in non-max mode in Agent mode to solve the remaining typescript errors.
So basically, someone must be sabotaging Cursor from within, because they can’t be that du*mb to mindlessly mess up into a garbage an once-upon-a-time excellent product.
Hi, as we use the feature request section to track unique requests, we often rename more popular requests to try to help users find them.
I don’t think the meaning of your request was lost here, but as GLM 4.5 isn’t yet available within Cursor, it makes sense to use the same post to allow people to vote on both adding the model to Cursor and using it behind Auto.
There are many feature requests we have already shipped, and new ones that will be shipped in the future too, so while we can’t guarantee all requests will be added, we do try our best to add the top ones, which are working on the future and stability of the editor.
@re365 Even in MAX mode, we do occasionally limit the maximum volume of context so that you don’t accidentally add a huge file and waste the whole 1 million tokens, but the Agent always has the ability to read whatever it wants and use the full context window.
While we did remove Manual mode, you can recreate it yourself with Custom Modes, if you want - the usage of the feature was very low, and not worth keeping in as a default.
R-I-D-I-C-U-L-O-U-S explanations no… the Custom Mode with only the Edit tool enabled will count each snippet/edit as a separate AI request… and with no tools enabled will just output the snippets without button/UI to apply them to the codebase.. IT IS NOT POSSIBLE TO RECREATE THE “MANUAL EDIT” BEHAVIOR IN THE CURRENT VERSION OF CURSOR
And fox MAX MODE I don’t know if you realize how du*mb what you said sound, the point is: IT IS NOT POSSIBLE TO RECREATE THE “MAX MODE” BEHAVIOR IN THE CURRENT VERSION OF CURSOR … do you understand what “NOT POSSIBLE” means? it means even if I had to click OK through 10 WARNING DIALOGS EACH TIME then I would still not be able to use the old “MAX MODE” behavior .. AND THAT IS DUMB, YOU ARE OVER-REACHING YOUR PURPOSE AS AN IDE, the content of the user’s intent with the provider’s machine learning business product IS BEYOND YOUR PURVIEW for you to interfere
