GPT Pro

I’m perfectly happy with 5.6 Sol for chat and 6.1 Sol for work. It’s you that create problems by removing the models I am already happy with.

Please make multi-bot architecture, similar to Grok Bot, available in ChatGPT Dots as soon as possible. This could bring the vast majority of users back to subscribing to ChatGPT! This should be the ChatGPT team's highest priority!

Your agent’s responses are responsible for 80% of the token burn. Why don’t you work on that?

GPT Pro has been really generous with its usage limits lately, but there’s one downside: way too many models to choose from. But with Codex, I still have to break down tasks myself and figure out which model to use for each one. I really hope you guys roll out an Auto mode soon, or at least let us pick a reasoning level like in Chat, so Codex can automatically plan out tasks and route them to the right models without us having to micromanage everything.