Codex Rate Limits Discussion Thread

Yes. OpenAI recommends this model as token-efficient (supposedly ideal for the PLUS plan) but using it leads you into a trap where you end up spending twice as much time and tokens simply because the model is so limited that it can only edit precisely specified lines and parameters. For other tasks it is useless. As a result, you end up having to use the GPT-5.4 model and burn through your weekly limit in a single day. I used to use up my limit in two days — I should have been happy while I could