Cursor vs. New ChatGPT, my ChatGPT expires in 2 days but

So I want:

-Code base wide queries.
-Uploading screen shots and other images
-web browsing
-Which has the higher maximum length for prompts and responses?

For me right now it’s neck-and-neck since being in an IDE and some of the other functionality including chat in terminal is very nice for Cursor.

But ChatGPT feels like there are more use-cases and of course NO LIMITS except the 50 I get every 3 hours which I have only reached once or twice.

Developers of Cursor, please feel free to do your best to convince me to buy Cursor. I want to hear about all the advantages for my use cases.

What I wish Cursor would do is for me to write a prompt say for a simple Spring Boot “Hello World” directory with one java file, ie. like “Write all the necessary files to compile and run this and organize the directories so it’s a Spring Boot app” or say I have one .cu Cuda file has a function that is used as an extern or external function in a .cpp driver file I’d want it to write the .cpp driver file, organize my directory, write the make file or cmake files to compile/build everything and then give me all the commands to run. Possible in 2025?

if this is your biggest concern, i would try using your own OpenAI key inside cursor. (your rate limit will be whatever OpenAI gives you…almost certainly high enough to program all day with Cursor & GPT-4)

1 Like

I tried and very quickly I’m at $1.80. That codebase wide query really eats up a lot of tokens right? Is there a page to see how that’s calculated? Is it basically sending each file in the codebase as context to the API?

Codebase-wide questions take the most relevant snippets from your codebase and put them in your prompt.