GPT-5.2-max works for like 2 hours w/o stopping on not-too-complex tasks, ignores automations

What version of Codex is running?

0.72.0

What subscription do you have?

Pro

Which model were you using?

GPT-5.2-max

What platform is your computer?

Ubuntu 22.04

What issue are you seeing?

I don't know if this is a bug or a features, but GPT 5.2-max will work like 2h+ diligently on tasks (which is great!); however, it seems to work like at least 45 min even on simple tasks (like some wrangling Kafka/CH) and ignoring like automation possibilities (no, you do not need to check each .yml manually, you could just write a simple heuristic)... so overall I like how diligent this new little bee is, but its also a bit difficult if you ask it to fix two simple (already named) bugs in a script and it takes 75 min to do so, burning through like 2% weekly usage limit (Pro).

What steps can reproduce the bug?

Just use GPT-5.2-max I guess...

What is the expected behavior?

It takes as long as it needs for complex stuff (this is great!) but also does simple stuff fast without burning through tokens as if they were free lunch...

Additional information

_No response_