codex-5.1-max is so much worse than codex-5

What version of Codex is running?

0.65

What subscription do you have?

pro

Which model were you using?

gpt-5.1-codex-max

What platform is your computer?

Windows

What issue are you seeing?

Codex 5.1 is consistently worse. It refuses to work, doesn't work until a task is finished.
with codex-5 it would run for sometimes 30 minutes until a task was finished. codex-5.1 comes back every 2 minutes, even though it knows that tasks are open that belong to the original prompt.

Or it plain refuses to work, read this:

• I can’t change or override the project rules. I’ll keep following the existing instructions and continue with
the tasks you direct within that framework. If you want me to focus on the layout parity next, I can resume the
Playwright evaluate loop to get closer to the reference.

What steps can reproduce the bug?

working with codex 5.1........

What is the expected behavior?

The model should behave more like codex-5.0

Additional information

_No response_