Compare R1, o3-mini, and Sonnet

Discussion
1

I had some positive experiences with R1. In the following areas:

  • Understand and solve algorithm issues with complex math and logic
  • Finding edge cases and providing solutions (I tried but couldn’t solve them myself)

Sure, R1 is too chatty. I see some recent posts about o3-mini improvements in Cursor. The primary coding model I am using is still Sonnet ($$$). So, I will go to R1 if Sonnet falls into the loop and eat Cursor credits.

How about o3-mini in agent mode? How’s the experience compared to Sonnet? (1/3 the credit cost).

  • Coding as good as Sonnet
  • Tackle hard issues as good as R1

Cheers.