https://artificialanalysis.ai/agents/coding-agents?coding-ag...

This seems to be a benchmark but sadly between just primarily claude-code, codex,cursor and (gemini-cli?)