aah, I can force it to load 5.5 by running it with `-m gpt-5.5`

hacks!

I'm curious where it will end up on the areana chart.. My guess is just below 4.7 opus thinking. ie #2.. maybe 3.. but it should be a jump up.. or maybe just my use cases.. I guess we'll see.

Modified by: PitViper on 2026-04-24 10:30:28 AM

Oh they better not start this...

I hope they fix this artifact issue on image gen 2 on chatgpt..

Deleted by: Fluchti

Hmm.. I had thought it would have scored higher. Gpt 5.5 High made it to #9

From what I've seen on reddit and my uses with it, it seems pretty much the same as 5.4.

same with opus 4.7 really not much over 4.6

So, I was monitoring my usage at work with claude enterprise.
I can do A LOT at home with 5x and opus.
At work, yesterday in about 4 hours, of some simple documentation work and updating a 12 repositories with new versions of some dependencies. No code changes just dependency updates and verifying builds and creating PRs. I spent about $200 of credits.

I didn't need to use opus for that, I shouldn't have but I did

Claude Code supports OTEL if you didn't know; can track quite a bit of data around your usage.

https://code.claude.com/docs/en/monitoring-usage

Modified by: swiftraccoon on 2026-05-01 07:27:11 PM

another useful thing is the status line
https://code.claude.com/docs/en/statusline

Modified by: swiftraccoon on 2026-05-01 07:46:05 PM