My current usage of Cursor
Unproofread notes
As I have been using Cursor a lot for coding these days, here is how I am using it currently:
- Gemini 2.5 Pro: for logic and complex tasks
- Claude 3.7 Sonnet: for the good looking UI
- o3 or MAX models: for debugging if stuck
Today, I saw a tweet from Cursor's founder Eric talking about having a similar setup, here's the exact tweet:
using o3 for research & planning and gemini 2.5 pro for implementation feels really good right now
However, I received a reply to one of my tweets and it said that according to Aider, using o3
for planning and then GPT-4.1
for coding scores the highest. But... from my experience, I didn't find GPT-4.1
to write code as good as Gemini 2.5 Pro
or Claude 3.7 Sonnet
.
- ← Previous
Web browsers telemetry – 2025 - Next →
Cursor is bringing a major shift
Comment via email