I’m thinking the same thing, Codex literally ruined the codebases that I experimented with it on.

Currently GPT just works much better, and so does Gemini but it's more expensive right now. Going through Opencode stats, their claim is that Gemini is the current best model followed by GPT 5.4 on their benchmarks, but the difference is slim.

My personal experience is best with GPT but it could be the specific kind of work I use it for which is heavy on maths and cpp (and some LISP).

OpenAI replaced its founding engineers with Meta PMs. The shift towards consumer engagement metrics and marketing is apparent.

You can believe whatever you want. I found claude unusable due to limits. Codex works very well for my use cases.