GPT 5.2 Codex in VS Code suddenly rapidly consuming credits and context (15 credits and 26% context for one simple CSS change).

What version of Codex is running?

0.77.0 (codex-cli)

What subscription do you have?

Plus

Which model were you using?

GPT 5.2

What platform is your computer?

WSL

What issue are you seeing?

Rapidly consumed usage. I am using GPT 5.2 in VS Code on WSL. I just asked it to move some text down a bit and it consumed 15 credits, and said it used 26% context. Why suddenly is it consuming so many credits when even 1 or 2 days ago it was so much more efficient.

Yesterday it says I used 13 credits. I did so much yesterday. That's 13 credits across SO many prompts. Today I used 15 credits on one very simple change. And a similar amount on other very simple changes. I've blown through so many credits today until I realized so much context was being used so quickly and then decided to go online and see how many credits were being consumed.

I am using high, not extra high, but I've always predominantly used high, including yesterday and previous days.

What steps can reproduce the bug?

Prompting for anything results in huge credit and context consumption.

What is the expected behavior?

Less credit and context usage is what's expected.

Additional information

_No response_