Token usage reporting in Codex to show Input token + Cache token in Token Usage

Resolved 💬 2 comments Opened Oct 4, 2025 by Chetansuri7 Closed Oct 16, 2025
💡 Likely answer: A maintainer (tibo-openai, collaborator) responded on this thread — see the highlighted reply below.

What feature would you like to see?

In Codex versions up to 0.41, the /status command displayed cache token usage separately, and the same breakdown was also shown after running /quit.

In the current version, the /status output now shows only total token usage (combined input and output), without separately listing cache tokens. While functional, this change makes it more difficult to calculate the precise cost of a task, since cache tokens are billed differently.

Example from current version:

Token usage: 7.49K total (7.38K input + 105 output)
Context window: 100% left (7.49K used / 272K)
Limits: data not available yet
Token usage: total=7,485 input=7,380 output=105

Request: Restore the separate cache token figure in /status and /quit outputs to improve cost calculation accuracy.

Additional information

_No response_

View original on GitHub ↗

2 Comments

tibo-openai collaborator · 10 months ago

This is an intentional change, marking as resolved but we might revisit in the future!

mkaalto · 9 months ago
This is an intentional change, marking as resolved but we might revisit in the future!

The change made the output less clear than previously. Could you describe, or link to the discussion why the change was made & what was the reasoning behind? Thank you