“I have Claude, ChatGPT and Cursor subs and I absolutely prefer Cursor as my daily driver for maintenance and developement. It often shits the bed for creative stuff, but the others are good enough for that.”
“I appreciate both, but not the middle ground that the Cursor IDE became. I’d rather have Cursor Agent as an app than a window, and Zed as a real no-AI IDE. At home it’s GPT & Claude plus Zed.”
“In this sub we always see people complaining about AI-powered IDEs like Cursor and Antigravity, switching to Claude Code and Codex (AI agents that’s don’t show diffs as an IDE and don’t have same methods of repo indexing)”
“No, but this subreddit is filled with shills, they seem to get paid to crap on cursor and grok, i like both but i dont post here either, use whats useful to you and ignore the noise”
“anyone who used twitter can tell you how it became a bloated piece of useless crap, full of bugs and issues, Elon is using the same playbook for cursor, jump ship or avoid using it while you can to avoid wasting your time getting familiar with it”
“Now that Grok 4.7 is out, Cursor is more and more losing the one thing that made it worth it. You used to be able to run it forever and never worry about limits.”
“I feel like going back to cursor. I don't like how I can only use Claude models. They kinda have the same pattern which is good for some stuff but not that good for others.”
“The nut of the problem for me is that Cursor is a middleman whose biggest competitors are the frontier labs themselves: Codex, Claude Code, etc. Plus a rapidly improving ecosystem of open-source tools like OpenCode.”
“If you love Cursor harness (I am) and accept to use Grok/Composer models pair with other frontier models, than yes. You can search CursorBench, on Cursor harness, Grok model is on pair with Fable-level.”
“cursor isn't locked to one company's models. you can run claude, gpt, or gemini under the same sub, so if a model gets dumb for a while (which happens to all of them) you just switch instead of being stuck waiting for openai to fix it.”
“i understand open code similar to claude code just that its multi model and more open but cursor being an agentic integrated IDE whats the cost cut they do to rug us ?”
“cursor is surprisingly cheap with the composer 2.5 model which is just as good in my opinion as the frontier models -- also cursors instant grep is prob my favorite”
“To me, nothing beats Cursor from purely the code aspect. Its integrations with the actual code are the best imo. Being able to have it callout lines in the chat window, clicking on those lines and getting a highlighted block in the editor is first class.”
“I feel confident and comfortable in Cursor and I can’t say the same for other tools. I’ve always used Cursor from the beginning and it’s difficult for me using something else.”
“* *"Cursor's slow requests feature got really painful after hitting the monthly quota limit. Switched over to Freebuff last week to finish up a web app project, and the response speed was pretty solid for a free setup."*”
“I got Cursor a few days ago, and surprisingly, I feel like I'm enjoying coding again. It feels like *I'm* the one doing things and the AI is assisting me, rather than me just supervising an agent doing everything.”
“Just use VS code with claude code or codex .... Cursor was amazing at the beginning with that inline suggestion that was superior to VS one.. but after claude code came no one is writing code by hand then why pay for cursor, like you can use their model yeah sure and got shitty result nothing can compare to codex/claude code .... GUYS DONT WASTE MONEY :D”
“also using the cursor but its good for moderate task and research kinda only when you use grok models, the moment you use anthropic models your monthly limit will hit within a hour or two max, thats pretty bad.”
“I’m currently using Cursor Pro, but lately it feels like the credits disappear almost instantly. Auto mode is also nowhere near as good as it was a few weeks ago, and the agent feels painfully slow when working on tasks.”
“I was using Cursor, but now I’ve switched to Claude. Cursor’s model limits and the fact that it was forcing me to use Grok were annoying, so I found Claude to be better compared to Cursor Composer and Grok.”
“Right now I'm running Cursor for daily edits/autocomplete and Claude Code for heavier agentic refactors. Codex is fine if you're already deep in OpenAI's ecosystem.”
“Jev has completely replaced ChatGPT, Codex, Cursor, Claude Code, iMessage, Spotify, Ford, CVS, and McDonalds for me. It really is the everything model.”
“Bro Last month cursor was 50% off and and ab addition discount (I think 50%) if you use the agent window instead of the IDE. So you can’t go by last month.”
“…that is why I canceled my cursor subscription (also because of non-transparency in usage pricing) and went back to VS code + Claude code in terminal and chat window (very similar to cursor UI)”
“I can't get off Cursor. But they're trying really hard to get me to leave with the defaulting to agents window and constantly switching my model to Grok away from Opus. But I keep using Cursor anyway because I still actually read/write code by hand and like the VS Code compatibility for the extensions.”
“newest-wins is predictable until clocks lie, which is basically what bit you — the cursor note was newer by timestamp but staler by causality. i've started keeping the conflict visible instead of resolving it: both notes stay, tagged, and the agent sees that they disagree.”
“I have tried codex and cursor and honestly the usage limits can be more annoying than the actual model differences. for bigger tasks I usually care more about how well it handles the whole repo without losing context halfway through”