Token Consumption

What version of Codex CLI is running?

Token Consumption

What subscription do you have?

business

Which model were you using?

5.4

What platform is your computer?

_No response_

What terminal emulator and version are you using (if applicable)?

_No response_

What issue are you seeing?

Hello OpenAI Support Team,

I am writing to report a billing issue caused by what appears to be a bug in Codex that has been affecting my account since approximately March 6, 2025.

Issue description:
Over the past 3 weeks, Codex has been consuming significantly more tokens than expected, even for simple tasks. Completion times for straightforward requests have reached up to 4 hours. Additionally, after the 'Context compacted' step, Codex repeats previously completed actions, triggering further unnecessary token usage. This creates a feedback loop that I was unable to interrupt without losing progress.

Due to this behavior, I was forced to purchase additional tokens totaling approximately $630 between March 6 and today (March 25, 2025).

1. A refund of $630 (or equivalent token credit) to my account for charges incurred due to this malfunction.
2. Investigation and resolution of the described token overconsumption loop.

Thank you for your assistance.

Best,
Adrian

What steps can reproduce the bug?

Uploaded thread: 019d0357-c758-7a43-bb93-2bc9e8f67461

What is the expected behavior?

_No response_

Additional information

_No response_