[ctbellmar]

I know it's been mentioned a few times, but worth repeating: these LLMs tend to do noticeably better in their own native environments. Claude (Opus or Sonnet) in Copilot != Claude in Claude Code. Same applies to Cursor, Windsurf, Augment, etc. This likely has a lot to do with context manipulation (and compression), which affects the resulting output. I imagine that GPT-5 likewise will do better in Codex vs 3rd party plugin/VS Code fork.

[fragmede]

The system prompts aren't shared either, and probably accounts for quite a bit of difference as well.