Limitations on the OpenAI o-series reasoning models on ChatGPT

It is corrected on the FAQ page, thanks!