Option to show model reasoning/thinking chain

[liraz747]

Describe the feature or problem you'd like to solve

It'd really be nice if we could have the option via some keyboard shortcut toggle to see thinking models' reasoning chain, like in the VS Code extension

Proposed solution

Some models, like gpt 5.2 on high reasoning can think for a very, very long time before doing anything, and it'd be nice if we could actually see it thinking and working rather than feeling like it's just not doing anything for 10 minutes sometimes, then runs a command- would also be useful to know to interrupt it when it has gotten things wrong, and really this is just a standard feature I think, supported in Gemini CLI and Claude Code and in Copilot's VS Code extension too

Example prompts or workflows

• User: "Debug why X isn't working"
• Model works without executing any tool for 10 minutes
• At this point the user would think the model isn't doing anything at all and probably get frustrated and just restart the CLI thinking something's broken, but if the thinking chain would be visible, we'd actually see it do something, and furthermore, how it thinks about actions as its about to do them, seeing the context of how it decides what action to do rather than just very short text saying "I'll edit X" with no info or just running tools without writing any text because all of the thinking is done in the reasoning chain.

Additional context

No response

[williammartin]

Hey @liraz747, is ctrl+t what you are looking for?

[liraz747]

@williammartin Thank you! I completely missed that

[williammartin]

No worries, there used to be a keybinding hint under the input prompt but I think it got lost last week. Should be fixed next release one way or another.