The main working models currently in use: Opus 4.5 - Enabling Cursor Plan Mode, including planning and implementation of large and complex features. Composer 1 - No Plan Mode required, small optimizations and minimal feature implementation, fast and quick, done in no time. GPT-5.1 Codex High - bugfix: I don't know if it's some kind of magic, but several tricky bugs were fixed using it. Just a few lines of code, and other models couldn't find the problem. Calm down, no more nonsense.
Loading thread detail
Fetching the original tweets from X for a clean reading view.
Hang tight—this usually only takes a few seconds.
