I’m experiencing unusually fast Codex usage exhaustion on ChatGPT Plus, and the behavior is very different from what I experienced before Previously, I could use higher reasoning levels such as Sol High for much longer coding sessions. Recently, even Sol Medium is exhausting my available usage after only a few short tasks.
My most recent example:
- Task 1: approximately 2 minutes
- Task 2: approximately 1 minute
- Task 3: approximately 1 minute
- Task 4: approximately 5 minutes
That is roughly 9 minutes of visible Codex work in total, and my usage allowance was exhausted before the fourth task was even completed.
I experienced a similar issue after another recent reset, where only a few tasks totaling around 29 minutes of visible work exhausted the entire usage window.
Has anyone else noticed a significant increase in Codex usage consumption recently?
It would also be helpful if OpenAI could clarify whether:
- model usage weighting has recently changed,
- Plus Codex limits have changed,
- usage metering has changed,
- or there is currently an issue causing unusually high usage consumption