Codex Rate Limits Discussion Thread

Update on Case #09851164: After the latest reset/window behavior, my Pro 20x quota still feels drastically lower than the previous Pro experience on the same account. This does not look like only a “reset timing UX” issue. The concerning part is that the visible burn rate still feels much faster than my actual active Codex work. From the user side, it even feels like existing local/cloud sessions, compaction, cached context, or rolling/periodic accounting may continue consuming quota when I am not actively starting new work. I am also seeing both Chinese and international/U.S. users on X/Twitter and in this community reporting similar Codex / Pro quota burn concerns. Can OpenAI please clarify whether Pro 20x quota/accounting, cached context/compaction, rolling-window behavior, model routing, or cloud/local session metering changed recently? This needs account-level quota/accounting review, not only a generic help-center explanation.