Option to show CoT instead of hiding by default

You said it hides by default, is there any way to view that?

@YasinHossain Control + T

I think for GPT models you get the CoT summaries still, but for GPT-OSS there aren't any so using that in codex is a very "quiet" experience, too quiet.

The new transcript mode is unusable. It is extremely difficult to keep track of it.

I liked the old style where you could see its thoughts and interrupt it to let it know its going off a red herring or not.

Not only is it enjoyable to read the chain of thought, it also allows the user to catch the model before it does something incorrectly and stop it from acting. With CoT hidden, you only notice the model deviate from your instructions once it starts to act, which is later than it could be.

@YasinHossain Control + T

For that to work, I needed to add these model reasoning stuff to my config.toml

karthidreamr@fedora:~/.codex$ cat config.toml 
model = "gpt-5"
model_reasoning_effort = "medium"

[projects."/home/karthidreamr/ai-chat"]
trust_level = "trusted"


model_reasoning_summary = "detailed"
model_verbosity = "high"

[projects."/home/karthidreamr"]
trust_level = "trusted"

<img width="1254" height="765" alt="Image" src="https://github.com/user-attachments/assets/a462aaeb-79f8-4b60-9802-31c2c3f6b1fa" />

https://developers.openai.com/codex/local-config#cli

As others have noted, you can use ctrl+T to toggle this feature.

useless