I’ve been paying for Cursor for a while, and I’ve also been paying for Cody. For code completion, I use Cursor, and for chat, I use both. The issue I’m facing is that I’m hitting the 500 request limit very quickly with Cursor, and there’s no information on the Cursor dashboard about when I’ll get the next 500 requests. The real issue, however, is that I don’t understand why I’m not encountering any limits with Cody, despite using it much more frequently. Cody is always reliable and fast.
I appreciate that Cursor is more integrated, and I prefer its user interface. However, I’m very dissatisfied with the request limits. In fact, I’m paying $20 for Cursor and $9 for Cody! I’m almost certain that if you don’t change something about this limitation, many users will switch to other alternatives.
I flew through my first 500 fast-premium requests very fast (within a few days).
But that was when i was just getting used to Cursor and how it worked.
The next 500 have lasted a lot longer, and I think that has mainly been due to selecting gpt-4o-mini for smaller tasks (when I thought the context wasn’t going to be ‘heavy’) because those requests aren’t counted as premium:
In regard to when the 500 requests ‘renew’, I am assuming that is just at the start of the next 1 month subscription period?
Below is a post showing what I see in when logged in at:
Don’t forget to use your 10 claude-3-opus per day (Roughly 300 in total per month). Those are premium but have to be used daily or you lose them.
500 claude-3.5-sonnet and gpt-4o uses
up to 300 claude-3-opus
unlimited gpt-4o-mini
I try to use gpt-4o-mini for smaller tasks and then switch to claude-3-opus when gpt-4o-mini is unable to complete the task. Then when I am out of claude-3-opus uses for the day for harder tasks i finally switch to claude-3.5-sonnet and gpt-4o for the harder tasks which count against our 500 limit. I spend most of my time using gpt-4o-mini.
Yeah, unlimited (for pro) gpt 4-mini is a game-changer for me. Actually, it’s really good, better than turbo according to the leaderboards: https://arena.lmsys.org/
I don’t know why devs don’t talk about it in changelogs or on the price page though. I hope it will stand as it is or I’ll have the same thoughts as the topic starter.
@Niko_Bellic, I agree that o-mini is very good for it’s price, but the main issue for me is, I pay 9USD for Cody and I have unlimited 3.5 usage with it, which is way better in my use-cases than o-mini
I’m in the same boat! I used Cody last month, but as you know, Cody has its own set of problems. Now, it really comes down to what’s more important for you: unlimited GPT-4 mini + 500 3.5 uses per month + the Cursor UI, or unlimited (oficially, I’ve heard some users hit limitations) 3.5 in Cody. I know the choice isn’t obvious, but I think I’ll give Cursor a try because I love the new composer and how it manages the codebase index. Cody hasn’t worked well for me in that regard.
I agree with you, Cursor UI is much better, at least for me. A user posted a project to export the chats, which I can call with a hotkey to keep my Obsidian in sync with the chats. This is also very useful. There are many nice things about the Cursor UI, and the autocomplete is great too! However, I would love to see an improvement in the request limit. It’s so annoying; I’ve been out of requests for many days now, so they lasted for such a short time. I didn’t like that at all. Having so few requests in a paid service is a bit of a dealbreaker for me.
(fwiw, I also burned through my fast-premium when I first signed up, but I use my models more thoughtfully now, friendly tip: gpt-4o-mini is a non-premium model, plus make sure to use your 10 daily opus and long-chat model requests, see here for more info).
The trick for me is to pay for cursor and use it for the IDE AI capabilities, and I pay also for Cody, I use it all day and never and issue or limitation, every now and then I ask my self I still want to pay for both or keep only Cody, so much cheaper and never a limitation, also very nice community,
Have been using it for about 6 months, however this billing cycle it either did not refresh for me/ran out suddenly in like 2 days however that is not indicative of my usage.
The usage history also doesn’t show my usage properly.
paying for Cursor for “tab” and long context messages
500 messages a month is a joke (2 days of work for me)
GPT mini is not really good
But thx God Cody is not expensive
My theory is that Cody can hold lower prices because of its enterprise clients. All individuals for them are just advertising. Maybe Cursor doesn’t have many.
I think the reason for the Cody price could be that.
I don’t know what to do. For now, I’ll pay for both, but it’s a little off for me. I’m almost exclusively using Cody for chat and Cursor for autocomplete, but Cody costs me $9 while Cursor costs me $20. There’s something weird about that pricing.
And I was using Codeium, never a warning about usage, and it’s 12$ for month payed Yearly.
