News for hackers. Stuff that matters.

A comment on GPT-5.5. Parent comment

Re:GPT-5.5 by andriy_koval on Thursday April 23, 2026 @03:41PM (#47880661)
what is your setup for drawing pelican? Do you ask model to check generated image, find issues and iterate over it which would demonstrate models real abilities?
[ Parent | on HN ]
Re:GPT-5.5 by simonw on Thursday April 23, 2026 @04:12PM (#47881153)
It's generally one-shot-only - whatever comes out the first time is what I go with.

I've been contemplating a more fair version where each model gets 3-5 attempts and then can select which rendered image is "best".

[ Parent | on HN ]
Re:GPT-5.5 by irthomasthomas on Thursday April 23, 2026 @04:19PM (#47881290)
Try llm-consortium with --judging-method rank
[ Parent | on HN ]
Re:GPT-5.5 by andriy_koval on Thursday April 23, 2026 @04:14PM (#47881202)
I think it will make results way better and more representative of model abilities..
[ Parent | on HN ]
Re:GPT-5.5 by simonw on Thursday April 23, 2026 @04:16PM (#47881232)
It would... but the test is inherently silly, so I'm still not sure if it's worth me investing that extra effort in it.
[ Parent | on HN ]