New usage caps for Plus users of Codex?
As the title implies, in the last week or so I have noticed a steep decline in the amount of usage that you can squeeze out of a 5h codex usage cap, or even a weekly one.
I have not changed the model, its speed nor its reasoning settings. And still, now a single code change prompt of 5.5 High at Standard speed consumed 16% of the 5h usage cap. And it's not like it refactored a whole project: just +158 -43 changes.
It's not that I can't lower the token usage by choosing different settings, it's that clearly something has changed in the background and as a consumer I feel OpenAI is following the shrinkflation trend in the shadows.
Is this the case? Or am I just an outlier?