\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

Codex 5.3 extra-high is so good its nuts

also way faster than 5.2
Blathering center
  02/06/26
what do you use it for
Claret aromatic codepig
  02/06/26
AI-related development. I just used it on extra high mode an...
Blathering center
  02/06/26
The SWE Bench Pro graph with number of tokens vs. accuracy i...
topaz yarmulke school
  02/06/26
I got into the alpha for the codex app This thing is amaz...
cordovan rehab clown
  02/06/26
its so 180. I have heard people complain that its an electro...
robot daddy
  02/10/26


Poast new message in this thread



Reply Favorite

Date: February 6th, 2026 12:25 AM
Author: Blathering center

also way faster than 5.2

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49650246)



Reply Favorite

Date: February 6th, 2026 12:25 AM
Author: Claret aromatic codepig

what do you use it for

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49650249)



Reply Favorite

Date: February 6th, 2026 12:31 AM
Author: Blathering center

AI-related development. I just used it on extra high mode and it edited (and created some new unit tests) 14 files in one shot, and fixed it to a quality level that would have taken like 12 rewrites with codex 5.2. plus it was like twice as fast.

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49650256)



Reply Favorite

Date: February 6th, 2026 1:37 AM
Author: topaz yarmulke school

The SWE Bench Pro graph with number of tokens vs. accuracy is interesting. At the limit, it converges on 5.2 codex max. Models are now becoming more token efficient, but not necessarily developing new capabilities?

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49650280)



Reply Favorite

Date: February 6th, 2026 1:48 AM
Author: cordovan rehab clown

I got into the alpha for the codex app

This thing is amazing

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49650289)



Reply Favorite

Date: February 10th, 2026 2:04 PM
Author: robot daddy

its so 180. I have heard people complain that its an electron app and that developers won't use it. but whats funny is I normally do almost everything headless but now that this app is out i just want to do everything on here. and speed and scale aren't really an issue. its fast enough and in terms of scale the credits are enough if I wanted to add more compute and more complex automation it would just get out of control and I'd have no idea what the hell is going on in my codebase. the credits on pro are enough. i could probably push it past the limits but i just would lose track of whats happening

(http://www.autoadmit.com/thread.php?thread_id=5831622&forum_id=2\u0026mark_id=5310752#49661179)