GPT 5.3 codex on codex and cli internal requests bug

What version of the Codex App are you using (From “About Codex” dialog)?

v0.144.1

What subscription do you have?

Pro

What platform is your computer?

Mac OS

What issue are you seeing?

There's a bug when using different models on same chat, When using a model not gpt 5.3 codex and have being used most of the context length and switch to gpt 5.3 codex it fails to summarize or continue, it throws a call tool called summary.reasoning error, and sometimes when trying to continue a chat when previously has been using other models with reasoning or thinking and then switch to gpt 5.3 codex , it have similar reasoning error output.

And there's a minor bug, that its when using from the start gpt 5.3 codex, depending on the tasks, it because of the cache token or the input token length its really bigger, like 200k or 300k and up of input token/cache token it gave context window error and throws to start again a clean new chat, it means fails to call the auto compact tool.

What steps can reproduce the bug?

1. Have a big long task chat and being used different reasoning model.
2. Have 80% or up context length being used or context length higher than the gpt 5.3 codex.
3. switch to gpt 5.3 codex
4. try to continue the task.
5. summary.reasoning error pop up.

What is the expected behavior?

1. auto compact
2. support reasoning.summary tool?

Additional information

_No response_