I just updated my Codex macOS app, which enables the new GPT-5.5 model. I've intentionally kept the speed to "Standard" to not burn through my tokens too fast.
After the latest update of the app, I notices that OpenAI changed the speed to Fast without asking, a change that burns up to 1.5x more tokens.
After the latest update of the app, I notices that OpenAI changed the speed to Fast without asking, a change that burns up to 1.5x more tokens.