Real limits converted to API-equivalent $ value for Claude Code, Codex, Copilot (twitter.com)

Yeah but different agents & models can still consume different pricing on different task. We clearly see Claude generates more output tokens / thinks more for same tasks tehn OpenAI