OpenAI Codex Team Responds to Unexpected Usage Limit Drain
OpenAI has issued a hard reset of usage limits for all Codex users after multiple reports surfaced about credits being consumed unexpectedly—even while users were idle or not actively running coding tasks.
The announcement came from Tibo Sottiaux of the OpenAI Codex team, who confirmed that engineers are actively investigating the issue and have already reset everyone’s usage quotas while the investigation continues.
The move comes after developers and power users noticed their Codex limits depleting much faster than expected, sparking discussions across the community about potential bugs involving background processes and automated subagents.
Hard Reset Applied to All Users
According to Tibo Sottiaux, the Codex team has performed a hard reset of everyone’s usage limits to immediately reduce the impact on affected users.
The team is currently reviewing internal logs as part of what OpenAI calls its internal “RESET week,” with engineers working together in a dedicated “war room” to identify the root cause.
For users who continue to experience unexpected limit consumption, OpenAI also promised that additional manual resets will be issued where necessary.
Users Report Credits Being Consumed While Idle
The reports gained attention after several users claimed their Codex usage limits continued decreasing despite not actively interacting with the service.
Community discussions pointed toward several possible causes, including:
- Automatic spawning of Codex subagents
- Background execution of coding tasks
- Certain plugins or experimental features
- Processes continuing after the primary task had finished
At this stage, OpenAI has not confirmed the exact cause, and these remain user observations while the investigation is ongoing.
Investigation Underway
The Codex engineering team says it is carefully analyzing system logs to determine why some accounts experienced unusually high usage.
While no technical explanation has been shared yet, the company appears to be treating the issue as a priority by resetting limits globally before publishing a permanent fix.
Developers relying on Codex for software development can continue monitoring their usage dashboards while OpenAI works toward identifying the underlying problem.
What This Means for Codex Users
For now, affected users should notice refreshed usage limits following the global reset.
If usage continues draining unusually fast, OpenAI has indicated that additional manual resets will be available as engineers continue their investigation.
The incident highlights the complexity of AI coding agents, especially as newer features such as autonomous subagents and background task execution become more capable. Ensuring accurate usage accounting will be increasingly important as AI development tools handle larger and longer-running workloads.
OpenAI is expected to provide further updates once the investigation identifies the source of the unexpected usage drains.
Keep yourself updated with all the latest AI news by reading our full coverage here.
Please follow us on our Facebook page and X account for all latest and breaking Windows and Microsoft related news.
