I paid for ChatGPT Pro, but for over a month I’ve never actually had access to the Pro model—only GPT-5.5 Mini. That’s unacceptable for a paid plan. Claude Code is at least transparent when access is restricted. OpenAI should be transparent too.
I’m paying $200/month for ChatGPT Pro and selecting GPT-5.6 Pro, but new chats sometimes say they’re running GPT-5.5 mini or GPT-5.5 Thinking. Is the model being silently downgraded, or is model self-identification unreliable? Please clarify. @OpenAI@Codex